[midend-tco]修复命名重复问题

Pass_ID_List.md

@@ -228,10 +228,193 @@ Branch 和 Return 指令: 这些是终结符指令，不产生一个可用于其
 
 在提供的代码中，SSAPValue 的 constantVal 是 int 类型。这使得浮点数常量传播变得复杂。对于浮点数相关的指令（kFAdd, kFMul, kFCmp, kFNeg, kFNot, kItoF, kFtoI 等），如果不能将浮点值准确地存储在 int 中，或者不能可靠地执行浮点运算，那么通常会保守地将结果设置为 Bottom。一个更完善的 SCCP 实现会使用 std::variant<int, float> 或独立的浮点常量存储来处理浮点数。
 
+## LoopSR循环归纳变量强度削弱 关于魔数计算的说明
+
+魔数除法的核心思想是：将除法转换为乘法和移位
+
+数学原理：x / d ≈ (x * m) >> (32 + s)
+
+m 是魔数 (magic number)
+s 是额外的移位量 (shift)
+>> 是算术右移
+
+2^(32+s) / d ≤ m < 2^(32+s) / d + 2^s / d
+
+cd /home/downright/Compiler_Opt/mysysy && python3 -c "
+# 真正的迭代原因：精度要求
+def explain_precision_requirement():
+    d = 10
+    
+    print('魔数算法需要找到精确的边界值：')
+    print('目标：2^p > d * (2^31 - r)，其中r是余数')
+    print()
+    
+    # 模拟我们算法的迭代过程
+    p = 31
+    two_p = 2**p
+    r = two_p % d  # 余数
+    m = two_p // d  # 商
+    
+    print(f'初始状态 (p={p}):')
+    print(f'  2^{p} = {two_p:,}')
+    print(f'  2^{p} / {d} = {m:,} 余 {r}')
+    print(f'  需要检查: 2^{p} > {d} * (2^31 - {r}) = {d * (2**31 - r):,}')
+    print(f'  条件满足: {two_p > d * (2**31 - r)}')
+    print()
+    
+    # 继续迭代直到条件满足
+    iteration = 0
+    while r != 0 and m < 2**31:
+        iteration += 1
+        p += 1
+        two_p *= 2
+        m *= 2
+        r *= 2
+        if r >= d:
+            m += 1
+            r -= d
+            
+        print(f'迭代 {iteration} (p={p}):')
+        print(f'  m = {m:,}, r = {r}')
+        print(f'  m < 2^31? {m < 2**31} (2^31 = {2**31:,})')
+        
+        if m >= 2**31:
+            print(f'  ✗ 魔数 {m:,} 超出32位范围!')
+            break
+        elif r == 0:
+            print(f'  ✓ 余数为0，找到精确解')
+            break
+        
+        if iteration > 10:  # 防止无限循环
+            print('  ... (限制迭代次数)')
+            break
+
+explain_precision_requirement()
+"
+
+问题在于精度要求。理论魔数需要非常精确，不能简单地用 2^32 / d。让我展示真正的迭代原因：
+
+cd /home/downright/Compiler_Opt/mysysy && python3 -c "
+# 真正的迭代原因：精度要求
+def explain_precision_requirement():
+    d = 10
+    
+    print('魔数算法需要找到精确的边界值：')
+    print('目标：2^p > d * (2^31 - r)，其中r是余数')
+    print()
+    
+    # 模拟我们算法的迭代过程
+    p = 31
+    two_p = 2**p
+    r = two_p % d  # 余数
+    m = two_p // d  # 商
+    
+    print(f'初始状态 (p={p}):')
+    print(f'  2^{p} = {two_p:,}')
+    print(f'  2^{p} / {d} = {m:,} 余 {r}')
+    print(f'  需要检查: 2^{p} > {d} * (2^31 - {r}) = {d * (2**31 - r):,}')
+    print(f'  条件满足: {two_p > d * (2**31 - r)}')
+    print()
+    
+    # 继续迭代直到条件满足
+    iteration = 0
+    while r != 0 and m < 2**31:
+        iteration += 1
+        p += 1
+        two_p *= 2
+        m *= 2
+        r *= 2
+        if r >= d:
+            m += 1
+            r -= d
+            
+        print(f'迭代 {iteration} (p={p}):')
+        print(f'  m = {m:,}, r = {r}')
+        print(f'  m < 2^31? {m < 2**31} (2^31 = {2**31:,})')
+        
+        if m >= 2**31:
+            print(f'  ✗ 魔数 {m:,} 超出32位范围!')
+            break
+        elif r == 0:
+            print(f'  ✓ 余数为0，找到精确解')
+            break
+        
+        if iteration > 10:  # 防止无限循环
+            print('  ... (限制迭代次数)')
+            break
+
+explain_precision_requirement()
+"
+
+
+cd /home/downright/Compiler_Opt/mysysy && python3 -c "
+# 检查初始状态是否会溢出
+def check_initial_overflow():
+    int32_max = 2**31 - 1
+    
+    print('检查不同除数的初始魔数是否会溢出:')
+    print(f'32位有符号整数最大值: {int32_max:,}')
+    print()
+    
+    # 检查各种除数
+    divisors = [2, 3, 5, 7, 9, 10, 11, 13, 17, 25, 100, 1000]
+    
+    for d in divisors:
+        # 初始魔数 = 2^31 / d
+        initial_magic = (2**31) // d
+        
+        overflow = initial_magic > int32_max
+        status = \"溢出\" if overflow else \"安全\"
+        
+        print(f'd={d:4d}: 初始魔数 = 2^31/{d} = {initial_magic:10,} [{status}]')
+        
+    print()
+    print('结论: 初始状态下魔数不会溢出，溢出发生在迭代过程中')
+
+check_initial_overflow()
+"
+
+总结
+迭代的必要性：
+
+不是为了避免初始溢出（初始状态安全）
+是为了找到最精确的魔数，减少舍入误差
+每次迭代提高一倍精度，但魔数也翻倍
+溢出发生时机：
+
+初始状态：2^31 / d 总是在32位范围内
+迭代过程：2^32 / d, 2^33 / d, ... 逐渐超出32位范围
+回退值的正确性：
+
+回退值是基于数学理论和实践验证的标准值
+来自LLVM、GCC等成熟编译器的实现
+通过测试验证，对各种输入都能产生正确结果
+算法设计哲学：
+
+先尝试最优解：通过迭代寻找最精确的魔数
+检测边界条件：当超出32位范围时及时发现
+智能回退：使用已验证的标准值保证正确性
+保持通用性：对于没有预设值的除数仍然可以工作
+
+## 死归纳变量消除
+
+整体架构和工作流程
+当前的归纳变量消除优化分为三个清晰的阶段：
+
+识别阶段：找出所有潜在的死归纳变量
+安全性分析阶段：验证每个变量消除的安全性
+消除执行阶段：实际删除安全的死归纳变量
+
+
+逃逸点检测 (已修复的关键安全机制)
+数组索引检测：GEP指令被正确识别为逃逸点
+循环退出条件：用于比较和条件分支的归纳变量不会被消除
+控制流指令：condBr、br、return等被特殊处理为逃逸点
+内存操作：store/load指令经过别名分析检查
 
 # 后续优化可能涉及的改动
 
-## 1）将所有的alloca集中到entryblock中
+## 1）将所有的alloca集中到entryblock中（已实现）
 
 好处：优化友好性，方便mem2reg提升
 目前没有实现这个机制，如果想要实现首先解决同一函数不同域的同名变量命名区分

script/runit-riscv64.sh

@@ -60,11 +60,7 @@ display_file_content() {
 # 清理临时文件的函数
 clean_tmp() {
     echo "正在清理临时目录: ${TMP_DIR}"
-    rm -rf "${TMP_DIR}"/*.s \
-           "${TMP_DIR}"/*_sysyc_riscv64 \
-           "${TMP_DIR}"/*_sysyc_riscv64.actual_out \
-           "${TMP_DIR}"/*_sysyc_riscv64.expected_stdout \
-           "${TMP_DIR}"/*_sysyc_riscv64.o
+    rm -rf "${TMP_DIR}"/*
     echo "清理完成。"
 }
 

script/runit-single.sh

@@ -2,64 +2,67 @@
 
 # runit-single.sh - 用于编译和测试单个或少量 SysY 程序的脚本
 # 模仿 runit.sh 的功能，但以具体文件路径作为输入。
+# 此脚本应该位于 mysysy/script/
+
+export ASAN_OPTIONS=detect_leaks=0
 
 # --- 配置区 ---
-# 请根据你的环境修改这些路径
-# 假设此脚本位于你的项目根目录或一个脚本目录中
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
-# 默认寻找项目根目录下的 build 和 lib
 BUILD_BIN_DIR="${SCRIPT_DIR}/../build/bin"
 LIB_DIR="${SCRIPT_DIR}/../lib"
-# 临时文件会存储在脚本所在目录的 tmp 子目录中
 TMP_DIR="${SCRIPT_DIR}/tmp"
 
 # 定义编译器和模拟器
 SYSYC="${BUILD_BIN_DIR}/sysyc"
+LLC_CMD="llc-19" # 新增
 GCC_RISCV64="riscv64-linux-gnu-gcc"
 QEMU_RISCV64="qemu-riscv64"
 
 # --- 初始化变量 ---
 EXECUTE_MODE=false
+IR_EXECUTE_MODE=false # 新增
 CLEAN_MODE=false
-SYSYC_TIMEOUT=10      # sysyc 编译超时 (秒)
-GCC_TIMEOUT=10        # gcc 编译超时 (秒)
-EXEC_TIMEOUT=5        # qemu 自动化执行超时 (秒)
-MAX_OUTPUT_LINES=50   # 对比失败时显示的最大行数
-SY_FILES=()           # 存储用户提供的 .sy 文件列表
+OPTIMIZE_FLAG=""
+SYSYC_TIMEOUT=30
+LLC_TIMEOUT=10 # 新增
+GCC_TIMEOUT=10
+EXEC_TIMEOUT=30
+MAX_OUTPUT_LINES=20
+SY_FILES=()
 PASSED_CASES=0
 FAILED_CASES_LIST=""
+INTERRUPTED=false # 新增
 
+# =================================================================
 # --- 函数定义 ---
+# =================================================================
 show_help() {
     echo "用法: $0 [文件1.sy] [文件2.sy] ... [选项]"
-    echo "编译并测试指定的 .sy 文件。"
-    echo ""
-    echo "如果找到对应的 .in/.out 文件，则进行自动化测试。否则，进入交互模式。"
+    echo "编译并测试指定的 .sy 文件。必须提供 -e 或 -eir 之一。"
     echo ""
     echo "选项:"
-    echo "  -e, --executable         编译为可执行文件并运行测试 (必须)。"
+    echo "  -e                       通过汇编运行测试 (sysyc -> gcc -> qemu)。"
+    echo "  -eir                     通过IR运行测试 (sysyc -> llc -> gcc -> qemu)。"
     echo "  -c, --clean              清理 tmp 临时目录下的所有文件。"
-    echo "  -sct N                   设置 sysyc 编译超时为 N 秒 (默认: 10)。"
+    echo "  -O1                      启用 sysyc 的 -O1 优化。"
+    echo "  -sct N                   设置 sysyc 编译超时为 N 秒 (默认: 30)。"
+    echo "  -lct N                   设置 llc-19 编译超时为 N 秒 (默认: 10)。"
     echo "  -gct N                   设置 gcc 交叉编译超时为 N 秒 (默认: 10)。"
-    echo "  -et N                    设置 qemu 自动化执行超时为 N 秒 (默认: 5)。"
-    echo "  -ml N, --max-lines N     当输出对比失败时，最多显示 N 行内容 (默认: 50)。"
+    echo "  -et N                    设置 qemu 自动化执行超时为 N 秒 (默认: 30)。"
+    echo "  -ml N, --max-lines N     当输出对比失败时，最多显示 N 行内容 (默认: 20)。"
     echo "  -h, --help               显示此帮助信息并退出。"
+    echo ""
+    echo "可在任何时候按 Ctrl+C 来中断测试并显示当前已完成的测例总结。"
 }
 
-# --- 新增功能: 显示文件内容并根据行数截断 ---
 display_file_content() {
     local file_path="$1"
     local title="$2"
     local max_lines="$3"
-
-    if [ ! -f "$file_path" ]; then
-        return
-    fi
-
+    if [ ! -f "$file_path" ]; then return; fi
     echo -e "$title"
     local line_count
     line_count=$(wc -l < "$file_path")
-    
     if [ "$line_count" -gt "$max_lines" ]; then
         head -n "$max_lines" "$file_path"
         echo -e "\e[33m[... 输出已截断，共 ${line_count} 行 ...]\e[0m"
@@ -68,51 +71,79 @@ display_file_content() {
     fi
 }
 
+# --- 新增：总结报告函数 ---
+print_summary() {
+    local total_cases=${#SY_FILES[@]}
+    echo ""
+    echo "======================================================================"
+    if [ "$INTERRUPTED" = true ]; then
+        echo -e "\e[33m测试被中断。正在汇总已完成的结果...\e[0m"
+    else
+        echo "所有测试完成"
+    fi
+
+    local failed_count
+    if [ -n "$FAILED_CASES_LIST" ]; then
+        failed_count=$(echo -e -n "${FAILED_CASES_LIST}" | wc -l)
+    else
+        failed_count=0
+    fi
+    local executed_count=$((PASSED_CASES + failed_count))
+
+    echo "测试结果: [通过: ${PASSED_CASES}, 失败: ${failed_count}, 已执行: ${executed_count}/${total_cases}]"
+
+    if [ -n "$FAILED_CASES_LIST" ]; then
+        echo ""
+        echo -e "\e[31m未通过的测例:\e[0m"
+        printf "%b" "${FAILED_CASES_LIST}"
+    fi
+    echo "======================================================================"
+
+    if [ "$failed_count" -gt 0 ]; then
+        exit 1
+    else
+        exit 0
+    fi
+}
+
+# --- 新增：SIGINT 信号处理函数 ---
+handle_sigint() {
+    INTERRUPTED=true
+    print_summary
+}
+
+# =================================================================
+# --- 主逻辑开始 ---
+# =================================================================
+
+# --- 新增：设置 trap 来捕获 SIGINT ---
+trap handle_sigint SIGINT
+
 # --- 参数解析 ---
-# 使用标准的 while 循环来健壮地处理任意顺序的参数
 while [[ "$#" -gt 0 ]]; do
     case "$1" in
-        -e|--executable)
-            EXECUTE_MODE=true
-            shift # 消耗选项
-            ;;
-        -c|--clean)
-            CLEAN_MODE=true
-            shift # 消耗选项
-            ;;
-        -sct)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then SYSYC_TIMEOUT="$2"; shift 2; else echo "错误: -sct 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -gct)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then GCC_TIMEOUT="$2"; shift 2; else echo "错误: -gct 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -et)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then EXEC_TIMEOUT="$2"; shift 2; else echo "错误: -et 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -ml|--max-lines)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then MAX_OUTPUT_LINES="$2"; shift 2; else echo "错误: --max-lines 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -h|--help)
-            show_help
-            exit 0
-            ;;
-        -*) # 未知选项
-            echo "未知选项: $1"
-            show_help
-            exit 1
-            ;;
-        *) # 其他参数被视为文件路径
+        -e|--executable) EXECUTE_MODE=true; shift ;;
+        -eir) IR_EXECUTE_MODE=true; shift ;; # 新增
+        -c|--clean) CLEAN_MODE=true; shift ;;
+        -O1) OPTIMIZE_FLAG="-O1"; shift ;;
+        -lct) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then LLC_TIMEOUT="$2"; shift 2; else echo "错误: -lct 需要一个正整数参数。" >&2; exit 1; fi ;; # 新增
+        -sct) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then SYSYC_TIMEOUT="$2"; shift 2; else echo "错误: -sct 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -gct) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then GCC_TIMEOUT="$2"; shift 2; else echo "错误: -gct 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -et) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then EXEC_TIMEOUT="$2"; shift 2; else echo "错误: -et 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -ml|--max-lines) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then MAX_OUTPUT_LINES="$2"; shift 2; else echo "错误: --max-lines 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -h|--help) show_help; exit 0 ;;
+        -*) echo "未知选项: $1"; show_help; exit 1 ;;
+        *)
             if [[ -f "$1" && "$1" == *.sy ]]; then
                 SY_FILES+=("$1")
             else
                 echo "警告: 无效文件或不是 .sy 文件，已忽略: $1"
             fi
-            shift # 消耗文件参数
+            shift
             ;;
     esac
 done
 
-
 if ${CLEAN_MODE}; then
   echo "检测到 -c/--clean 选项，正在清空 ${TMP_DIR}..."
   if [ -d "${TMP_DIR}" ]; then
@@ -121,19 +152,22 @@ if ${CLEAN_MODE}; then
   else
     echo "临时目录 ${TMP_DIR} 不存在，无需清理。"
   fi
-  
-  if [ ${#SY_FILES[@]} -eq 0 ] && ! ${EXECUTE_MODE}; then
+  if [ ${#SY_FILES[@]} -eq 0 ] && ! ${EXECUTE_MODE} && ! ${IR_EXECUTE_MODE}; then
     exit 0
   fi
 fi
 
-# --- 主逻辑开始 ---
-if ! ${EXECUTE_MODE}; then
-    echo "错误: 请提供 -e 或 --executable 选项来运行测试。"
+if ! ${EXECUTE_MODE} && ! ${IR_EXECUTE_MODE}; then
+    echo "错误: 请提供 -e 或 -eir 选项来运行测试。"
     show_help
     exit 1
 fi
 
+if ${EXECUTE_MODE} && ${IR_EXECUTE_MODE}; then
+    echo -e "\e[31m错误: -e 和 -eir 选项不能同时使用。\e[0m" >&2
+    exit 1
+fi
+
 if [ ${#SY_FILES[@]} -eq 0 ]; then
     echo "错误: 未提供任何 .sy 文件作为输入。"
     show_help
@@ -144,18 +178,18 @@ mkdir -p "${TMP_DIR}"
 TOTAL_CASES=${#SY_FILES[@]}
 
 echo "SysY 单例测试运行器启动..."
-echo "超时设置: sysyc=${SYSYC_TIMEOUT}s, gcc=${GCC_TIMEOUT}s, qemu=${EXEC_TIMEOUT}s"
-echo "失败输出最大行数: ${MAX_OUTPUT_LINES}"
+if [ -n "$OPTIMIZE_FLAG" ]; then echo "优化等级: ${OPTIMIZE_FLAG}"; fi
+echo "超时设置: sysyc=${SYSYC_TIMEOUT}s, llc=${LLC_TIMEOUT}s, gcc=${GCC_TIMEOUT}s, qemu=${EXEC_TIMEOUT}s"
 echo ""
 
 for sy_file in "${SY_FILES[@]}"; do
     is_passed=1
+    compilation_ok=1
     base_name=$(basename "${sy_file}" .sy)
     source_dir=$(dirname "${sy_file}")
 
-    ir_file="${TMP_DIR}/${base_name}_sysyc_riscv64.ll"
+    ir_file="${TMP_DIR}/${base_name}.ll"
     assembly_file="${TMP_DIR}/${base_name}.s"
-    assembly_debug_file="${TMP_DIR}/${base_name}_d.s"
     executable_file="${TMP_DIR}/${base_name}"
     input_file="${source_dir}/${base_name}.in"
     output_reference_file="${source_dir}/${base_name}.out"
@@ -164,47 +198,39 @@ for sy_file in "${SY_FILES[@]}"; do
     echo "======================================================================"
     echo "正在处理: ${sy_file}"
 
-    # --- 本次修改点: 拷贝源文件到 tmp 目录 ---
-    echo "  拷贝源文件到 ${TMP_DIR}..."
-    cp "${sy_file}" "${TMP_DIR}/$(basename "${sy_file}")"
-    if [ -f "${input_file}" ]; then
-        cp "${input_file}" "${TMP_DIR}/$(basename "${input_file}")"
-    fi
-    if [ -f "${output_reference_file}" ]; then
-        cp "${output_reference_file}" "${TMP_DIR}/$(basename "${output_reference_file}")"
-    fi
+    # --- 编译阶段 ---
+    if ${IR_EXECUTE_MODE}; then
+        # 路径1: sysyc -> llc -> gcc
+        echo "  [1/3] 使用 sysyc 编译为 IR (超时 ${SYSYC_TIMEOUT}s)..."
+        timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -s ir "${sy_file}" ${OPTIMIZE_FLAG} -o "${ir_file}"
+        if [ $? -ne 0 ]; then echo -e "\e[31m错误: SysY (IR) 编译失败或超时。\e[0m"; compilation_ok=0; fi
 
-    # 步骤 1: sysyc 编译
-    echo "  使用 sysyc 编译 (超时 ${SYSYC_TIMEOUT}s)..."
-    timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -s ir "${sy_file}" > "${ir_file}"
-    SYSYC_STATUS=$?
-    if [ $SYSYC_STATUS -eq 124 ]; then
-        echo -e "\e[31m错误: SysY 编译 ${sy_file} IR超时\e[0m"
-        is_passed=0
-    elif [ $SYSYC_STATUS -ne 0 ]; then
-        echo -e "\e[31m错误: SysY 编译 ${sy_file} IR失败，退出码: ${SYSYC_STATUS}\e[0m"
-        is_passed=0
-    fi
-    timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -S "${sy_file}" -o "${assembly_file}"
-    if [ $? -ne 0 ]; then
-        echo -e "\e[31m错误: SysY 编译失败或超时。\e[0m"
-        is_passed=0
-    fi
-    # timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -s asmd "${sy_file}" > "${assembly_debug_file}" 2>&1
+        if [ "$compilation_ok" -eq 1 ]; then
+            echo "  [2/3] 使用 llc 编译为汇编 (超时 ${LLC_TIMEOUT}s)..."
+            timeout -s KILL ${LLC_TIMEOUT} "${LLC_CMD}" -march=riscv64 -mcpu=generic-rv64 -mattr=+m,+a,+f,+d,+c -filetype=asm "${ir_file}" -o "${assembly_file}"
+            if [ $? -ne 0 ]; then echo -e "\e[31m错误: llc 编译失败或超时。\e[0m"; compilation_ok=0; fi
+        fi
 
-    # 步骤 2: GCC 编译
-    if [ "$is_passed" -eq 1 ]; then
-        echo "  使用 gcc 编译 (超时 ${GCC_TIMEOUT}s)..."
-        timeout -s KILL ${GCC_TIMEOUT} "${GCC_RISCV64}" "${assembly_file}" -o "${executable_file}" -L"${LIB_DIR}" -lsysy_riscv -static
-        if [ $? -ne 0 ]; then
-            echo -e "\e[31m错误: GCC 编译失败或超时。\e[0m"
-            is_passed=0
+        if [ "$compilation_ok" -eq 1 ]; then
+            echo "  [3/3] 使用 gcc 编译 (超时 ${GCC_TIMEOUT}s)..."
+            timeout -s KILL ${GCC_TIMEOUT} "${GCC_RISCV64}" "${assembly_file}" -o "${executable_file}" -L"${LIB_DIR}" -lsysy_riscv -static
+            if [ $? -ne 0 ]; then echo -e "\e[31m错误: GCC 编译失败或超时。\e[0m"; compilation_ok=0; fi
+        fi
+    else # EXECUTE_MODE
+        # 路径2: sysyc -> gcc
+        echo "  [1/2] 使用 sysyc 编译为汇编 (超时 ${SYSYC_TIMEOUT}s)..."
+        timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -S "${sy_file}" ${OPTIMIZE_FLAG} -o "${assembly_file}"
+        if [ $? -ne 0 ]; then echo -e "\e[31m错误: SysY (汇编) 编译失败或超时。\e[0m"; compilation_ok=0; fi
+
+        if [ "$compilation_ok" -eq 1 ]; then
+            echo "  [2/2] 使用 gcc 编译 (超时 ${GCC_TIMEOUT}s)..."
+            timeout -s KILL ${GCC_TIMEOUT} "${GCC_RISCV64}" "${assembly_file}" -o "${executable_file}" -L"${LIB_DIR}" -lsysy_riscv -static
+            if [ $? -ne 0 ]; then echo -e "\e[31m错误: GCC 编译失败或超时。\e[0m"; compilation_ok=0; fi
         fi
     fi
 
-    # 步骤 3: 执行与测试
-    if [ "$is_passed" -eq 1 ]; then
-        # 检查是自动化测试还是交互模式
+    # --- 执行与测试阶段 (公共逻辑) ---
+    if [ "$compilation_ok" -eq 1 ]; then
         if [ -f "${input_file}" ] || [ -f "${output_reference_file}" ]; then
             # --- 自动化测试模式 ---
             echo "  检测到 .in/.out 文件，进入自动化测试模式..."
@@ -227,24 +253,26 @@ for sy_file in "${SY_FILES[@]}"; do
                         EXPECTED_RETURN_CODE="$LAST_LINE_TRIMMED"
                         EXPECTED_STDOUT_FILE="${TMP_DIR}/${base_name}.expected_stdout"
                         head -n -1 "${output_reference_file}" > "${EXPECTED_STDOUT_FILE}"
-                        if [ "$ACTUAL_RETURN_CODE" -ne "$EXPECTED_RETURN_CODE" ]; then echo -e "\e[31m  返回码测试失败: 期望 ${EXPECTED_RETURN_CODE}, 实际 ${ACTUAL_RETURN_CODE}\e[0m"; is_passed=0; fi
                         
+                        ret_ok=1
+                        if [ "$ACTUAL_RETURN_CODE" -ne "$EXPECTED_RETURN_CODE" ]; then echo -e "\e[31m  返回码测试失败: 期望 ${EXPECTED_RETURN_CODE}, 实际 ${ACTUAL_RETURN_CODE}\e[0m"; ret_ok=0; fi
+                        
+                        out_ok=1
                         if ! diff -q <(tr -d '[:space:]' < "${output_actual_file}") <(tr -d '[:space:]' < "${EXPECTED_STDOUT_FILE}") >/dev/null 2>&1; then
-                            echo -e "\e[31m  标准输出测试失败。\e[0m"
-                            is_passed=0
+                            echo -e "\e[31m  标准输出测试失败。\e[0m"; out_ok=0
                             display_file_content "${EXPECTED_STDOUT_FILE}" "    \e[36m--- 期望输出 ---\e[0m" "${MAX_OUTPUT_LINES}"
                             display_file_content "${output_actual_file}" "    \e[36m--- 实际输出 ---\e[0m" "${MAX_OUTPUT_LINES}"
-                            echo -e "    \e[36m----------------\e[0m"
                         fi
+
+                        if [ "$ret_ok" -eq 1 ] && [ "$out_ok" -eq 1 ]; then echo -e "\e[32m  返回码与标准输出测试成功。\e[0m"; else is_passed=0; fi
+
                     else
                         if diff -q <(tr -d '[:space:]' < "${output_actual_file}") <(tr -d '[:space:]' < "${output_reference_file}") >/dev/null 2>&1; then
                             echo -e "\e[32m  标准输出测试成功。\e[0m"
                         else
-                            echo -e "\e[31m  标准输出测试失败。\e[0m"
-                            is_passed=0
+                            echo -e "\e[31m  标准输出测试失败。\e[0m"; is_passed=0
                             display_file_content "${output_reference_file}" "    \e[36m--- 期望输出 ---\e[0m" "${MAX_OUTPUT_LINES}"
                             display_file_content "${output_actual_file}" "    \e[36m--- 实际输出 ---\e[0m" "${MAX_OUTPUT_LINES}"
-                            echo -e "    \e[36m----------------\e[0m"
                         fi
                     fi
                 else
@@ -253,20 +281,16 @@ for sy_file in "${SY_FILES[@]}"; do
             fi
         else
             # --- 交互模式 ---
-            echo -e "\e[33m"
-            echo "  **********************************************************"
-            echo "  ** 未找到 .in 或 .out 文件，进入交互模式。             **"
-            echo "  ** 程序即将运行，你可以直接在终端中输入。             **"
-            echo "  ** 按下 Ctrl+D (EOF) 或以其他方式结束程序以继续。 **"
-            echo "  **********************************************************"
-            echo -e "\e[0m"
+            echo -e "\e[33m\n  未找到 .in 或 .out 文件，进入交互模式...\e[0m"
             "${QEMU_RISCV64}" "${executable_file}"
             INTERACTIVE_RET_CODE=$?
-            echo -e "\e[33m\n  交互模式执行完毕，程序返回码: ${INTERACTIVE_RET_CODE}\e[0m"
-            echo "  注意: 交互模式的结果未经验证。"
+            echo -e "\e[33m\n  交互模式执行完毕，程序返回码: ${INTERACTIVE_RET_CODE} (此结果未经验证)\e[0m"
         fi
+    else
+      is_passed=0
     fi
 
+    # --- 状态总结 ---
     if [ "$is_passed" -eq 1 ]; then
         echo -e "\e[32m状态: 通过\e[0m"
         ((PASSED_CASES++))
@@ -277,20 +301,4 @@ for sy_file in "${SY_FILES[@]}"; do
 done
 
 # --- 打印最终总结 ---
-echo "======================================================================"
-echo "所有测试完成"
-echo "测试通过率: [${PASSED_CASES}/${TOTAL_CASES}]"
-
-if [ -n "$FAILED_CASES_LIST" ]; then
-    echo ""
-    echo -e "\e[31m未通过的测例:\e[0m"
-    echo -e "${FAILED_CASES_LIST}"
-fi
-
-echo "======================================================================"
-
-if [ "$PASSED_CASES" -eq "$TOTAL_CASES" ]; then
-    exit 0
-else
-    exit 1
-fi
+print_summary

script/runit.sh

@@ -1,31 +1,41 @@
 #!/bin/bash
 
 # runit.sh - 用于编译和测试 SysY 程序的脚本
-# 此脚本应该位于 mysysy/test_script/
+# 此脚本应该位于 mysysy/script/
+
+export ASAN_OPTIONS=detect_leaks=0
 
 # 定义相对于脚本位置的目录
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
 TESTDATA_DIR="${SCRIPT_DIR}/../testdata"
 BUILD_BIN_DIR="${SCRIPT_DIR}/../build/bin"
 LIB_DIR="${SCRIPT_DIR}/../lib"
-# TMP_DIR="${SCRIPT_DIR}/tmp"
 TMP_DIR="${SCRIPT_DIR}/tmp"
 
 # 定义编译器和模拟器
 SYSYC="${BUILD_BIN_DIR}/sysyc"
+LLC_CMD="llc-19"
 GCC_RISCV64="riscv64-linux-gnu-gcc"
 QEMU_RISCV64="qemu-riscv64"
 
-# --- 新增功能: 初始化变量 ---
+# --- 状态变量 ---
 EXECUTE_MODE=false
-SYSYC_TIMEOUT=10      # sysyc 编译超时 (秒)
-GCC_TIMEOUT=10        # gcc 编译超时 (秒)
-EXEC_TIMEOUT=5        # qemu 执行超时 (秒)
-MAX_OUTPUT_LINES=50   # 对比失败时显示的最大行数
-TEST_SETS=()          # 用于存储要运行的测试集
+IR_EXECUTE_MODE=false
+OPTIMIZE_FLAG=""
+SYSYC_TIMEOUT=30
+LLC_TIMEOUT=10
+GCC_TIMEOUT=10
+EXEC_TIMEOUT=30
+MAX_OUTPUT_LINES=20
+TEST_SETS=()
 TOTAL_CASES=0
 PASSED_CASES=0
-FAILED_CASES_LIST=""  # 用于存储未通过的测例列表
+FAILED_CASES_LIST=""
+INTERRUPTED=false # 新增：用于标记是否被中断
+
+# =================================================================
+# --- 函数定义 ---
+# =================================================================
 
 # 显示帮助信息的函数
 show_help() {
@@ -33,30 +43,32 @@ show_help() {
     echo "此脚本用于按文件名前缀数字升序编译和测试 .sy 文件。"
     echo ""
     echo "选项:"
-    echo "  -e, --executable         编译为可执行文件并运行测试。"
+    echo "  -e, --executable         编译为汇编并运行测试 (sysyc -> gcc -> qemu)。"
+    echo "  -eir                     通过IR编译为可执行文件并运行测试 (sysyc -> llc -> gcc -> qemu)。"
     echo "  -c, --clean              清理 'tmp' 目录下的所有生成文件。"
+    echo "  -O1                      启用 sysyc 的 -O1 优化。"
     echo "  -set [f|h|p|all]...    指定要运行的测试集 (functional, h_functional, performance)。可多选，默认为 all。"
-    echo "  -sct N                   设置 sysyc 编译超时为 N 秒 (默认: 10)。"
+    echo "  -sct N                   设置 sysyc 编译超时为 N 秒 (默认: 30)。"
+    echo "  -lct N                   设置 llc-19 编译超时为 N 秒 (默认: 10)。"
     echo "  -gct N                   设置 gcc 交叉编译超时为 N 秒 (默认: 10)。"
-    echo "  -et N                    设置 qemu 执行超时为 N 秒 (默认: 5)。"
-    echo "  -ml N, --max-lines N     当输出对比失败时，最多显示 N 行内容 (默认: 50)。"
+    echo "  -et N                    设置 qemu 执行超时为 N 秒 (默认: 30)。"
+    echo "  -ml N, --max-lines N     当输出对比失败时，最多显示 N 行内容 (默认: 20)。"
     echo "  -h, --help               显示此帮助信息并退出。"
+    echo ""
+    echo "注意: 默认行为 (无 -e 或 -eir) 是将 .sy 文件同时编译为 .s (汇编) 和 .ll (IR)，不执行。"
+    echo "      可在任何时候按 Ctrl+C 来中断测试并显示当前已完成的测例总结。"
 }
 
+
 # 显示文件内容并根据行数截断的函数
 display_file_content() {
     local file_path="$1"
     local title="$2"
     local max_lines="$3"
-
-    if [ ! -f "$file_path" ]; then
-        return
-    fi
-
+    if [ ! -f "$file_path" ]; then return; fi
     echo -e "$title"
     local line_count
     line_count=$(wc -l < "$file_path")
-    
     if [ "$line_count" -gt "$max_lines" ]; then
         head -n "$max_lines" "$file_path"
         echo -e "\e[33m[... 输出已截断，共 ${line_count} 行 ...]\e[0m"
@@ -71,61 +83,90 @@ clean_tmp() {
     rm -rf "${TMP_DIR}"/*
 }
 
-# 如果临时目录不存在，则创建它
+# --- 新增：总结报告函数 ---
+print_summary() {
+    echo "" # 确保从新的一行开始
+    echo "========================================"
+    if [ "$INTERRUPTED" = true ]; then
+        echo -e "\e[33m测试被中断。正在汇总已完成的结果...\e[0m"
+    else
+        echo "测试完成"
+    fi
+
+    local failed_count
+    if [ -n "$FAILED_CASES_LIST" ]; then
+        # `wc -l` 计算由换行符分隔的列表项数
+        failed_count=$(echo -e -n "${FAILED_CASES_LIST}" | wc -l)
+    else
+        failed_count=0
+    fi
+    local executed_count=$((PASSED_CASES + failed_count))
+
+    echo "测试结果: [通过: ${PASSED_CASES}, 失败: ${failed_count}, 已执行: ${executed_count}/${TOTAL_CASES}]"
+
+    if [ -n "$FAILED_CASES_LIST" ]; then
+        echo ""
+        echo -e "\e[31m未通过的测例:\e[0m"
+        # 使用 printf 保证原样输出
+        printf "%b" "${FAILED_CASES_LIST}"
+    fi
+
+    echo "========================================"
+
+    if [ "$failed_count" -gt 0 ]; then
+        exit 1
+    else
+        exit 0
+    fi
+}
+
+# --- 新增：SIGINT 信号处理函数 ---
+handle_sigint() {
+    INTERRUPTED=true
+    print_summary
+}
+
+# =================================================================
+# --- 主逻辑开始 ---
+# =================================================================
+
+# --- 新增：设置 trap 来捕获 SIGINT ---
+trap handle_sigint SIGINT
+
 mkdir -p "${TMP_DIR}"
 
 # 解析命令行参数
 while [[ "$#" -gt 0 ]]; do
     case "$1" in
-        -e|--executable)
-            EXECUTE_MODE=true
-            shift
-            ;;
-        -c|--clean)
-            clean_tmp
-            exit 0
-            ;;
+        -e|--executable) EXECUTE_MODE=true; shift ;;
+        -eir) IR_EXECUTE_MODE=true; shift ;;
+        -c|--clean) clean_tmp; exit 0 ;;
+        -O1) OPTIMIZE_FLAG="-O1"; shift ;;
         -set)
-            shift # 移过 '-set'
-            # 消耗所有后续参数直到遇到下一个选项
-            while [[ "$#" -gt 0 && ! "$1" =~ ^- ]]; do
-                TEST_SETS+=("$1")
-                shift
-            done
-            ;;
-        -sct)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then SYSYC_TIMEOUT="$2"; shift 2; else echo "错误: -sct 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -gct)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then GCC_TIMEOUT="$2"; shift 2; else echo "错误: -gct 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -et)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then EXEC_TIMEOUT="$2"; shift 2; else echo "错误: -et 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -ml|--max-lines)
-            if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then MAX_OUTPUT_LINES="$2"; shift 2; else echo "错误: --max-lines 需要一个正整数参数。" >&2; exit 1; fi
-            ;;
-        -h|--help)
-            show_help
-            exit 0
-            ;;
-        *)
-            echo "未知选项: $1"
-            show_help
-            exit 1
+            shift
+            while [[ "$#" -gt 0 && ! "$1" =~ ^- ]]; do TEST_SETS+=("$1"); shift; done
             ;;
+        -sct) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then SYSYC_TIMEOUT="$2"; shift 2; else echo "错误: -sct 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -lct) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then LLC_TIMEOUT="$2"; shift 2; else echo "错误: -lct 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -gct) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then GCC_TIMEOUT="$2"; shift 2; else echo "错误: -gct 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -et) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then EXEC_TIMEOUT="$2"; shift 2; else echo "错误: -et 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -ml|--max-lines) if [[ -n "$2" && "$2" =~ ^[0-9]+$ ]]; then MAX_OUTPUT_LINES="$2"; shift 2; else echo "错误: --max-lines 需要一个正整数参数。" >&2; exit 1; fi ;;
+        -h|--help) show_help; exit 0 ;;
+        *) echo "未知选项: $1"; show_help; exit 1 ;;
     esac
 done
 
-# --- 本次修改点: 根据 -set 参数构建查找路径 ---
+if ${EXECUTE_MODE} && ${IR_EXECUTE_MODE}; then
+    echo -e "\e[31m错误: -e 和 -eir 选项不能同时使用。\e[0m" >&2
+    exit 1
+fi
+
 declare -A SET_MAP
 SET_MAP[f]="functional"
 SET_MAP[h]="h_functional"
 SET_MAP[p]="performance"
 
 SEARCH_PATHS=()
-
-# 如果未指定测试集，或指定了 'all'，则搜索所有目录
 if [ ${#TEST_SETS[@]} -eq 0 ] || [[ " ${TEST_SETS[@]} " =~ " all " ]]; then
     SEARCH_PATHS+=("${TESTDATA_DIR}")
 else
@@ -138,23 +179,34 @@ else
     done
 fi
 
-# 如果没有有效的搜索路径，则退出
 if [ ${#SEARCH_PATHS[@]} -eq 0 ]; then
     echo -e "\e[31m错误: 没有找到有效的测试集目录，测试中止。\e[0m"
     exit 1
 fi
 
 echo "SysY 测试运行器启动..."
+if [ -n "$OPTIMIZE_FLAG" ]; then echo "优化等级: ${OPTIMIZE_FLAG}"; fi
 echo "输入目录: ${SEARCH_PATHS[@]}"
 echo "临时目录: ${TMP_DIR}"
-echo "执行模式: ${EXECUTE_MODE}"
-if ${EXECUTE_MODE}; then
-    echo "超时设置: sysyc=${SYSYC_TIMEOUT}s, gcc=${GCC_TIMEOUT}s, qemu=${EXEC_TIMEOUT}s"
+
+RUN_MODE_INFO=""
+if ${IR_EXECUTE_MODE}; then
+    RUN_MODE_INFO="IR执行模式 (-eir)"
+    TIMEOUT_INFO="超时设置: sysyc=${SYSYC_TIMEOUT}s, llc=${LLC_TIMEOUT}s, gcc=${GCC_TIMEOUT}s, qemu=${EXEC_TIMEOUT}s"
+elif ${EXECUTE_MODE}; then
+    RUN_MODE_INFO="直接执行模式 (-e)"
+    TIMEOUT_INFO="超时设置: sysyc=${SYSYC_TIMEOUT}s, gcc=${GCC_TIMEOUT}s, qemu=${EXEC_TIMEOUT}s"
+else
+    RUN_MODE_INFO="编译模式 (默认)"
+    TIMEOUT_INFO="超时设置: sysyc=${SYSYC_TIMEOUT}s"
+fi
+echo "运行模式: ${RUN_MODE_INFO}"
+echo "${TIMEOUT_INFO}"
+if ${EXECUTE_MODE} || ${IR_EXECUTE_MODE}; then
     echo "失败输出最大行数: ${MAX_OUTPUT_LINES}"
 fi
 echo ""
 
-# 使用构建好的路径查找 .sy 文件并排序
 sy_files=$(find "${SEARCH_PATHS[@]}" -name "*.sy" | sort -V)
 if [ -z "$sy_files" ]; then
     echo "在指定目录中未找到任何 .sy 文件。"
@@ -162,139 +214,229 @@ if [ -z "$sy_files" ]; then
 fi
 TOTAL_CASES=$(echo "$sy_files" | wc -w)
 
-# --- 修复: 使用 here-string (<<<) 代替管道 (|) 来避免子 shell 问题 ---
 while IFS= read -r sy_file; do
-    is_passed=1 # 1 表示通过, 0 表示失败
+    is_passed=0 # 0 表示失败, 1 表示通过
 
     relative_path_no_ext=$(realpath --relative-to="${TESTDATA_DIR}" "${sy_file%.*}")
     output_base_name=$(echo "${relative_path_no_ext}" | tr '/' '_')
 
-    assembly_file="${TMP_DIR}/${output_base_name}_sysyc_riscv64.s"
-    executable_file="${TMP_DIR}/${output_base_name}_sysyc_riscv64"
+    assembly_file_S="${TMP_DIR}/${output_base_name}_sysyc_S.s"
+    executable_file_S="${TMP_DIR}/${output_base_name}_sysyc_S"
+    output_actual_file_S="${TMP_DIR}/${output_base_name}_sysyc_S.actual_out"
+
+    ir_file="${TMP_DIR}/${output_base_name}_sysyc_ir.ll"
+    assembly_file_from_ir="${TMP_DIR}/${output_base_name}_from_ir.s"
+    executable_file_from_ir="${TMP_DIR}/${output_base_name}_from_ir"
+    output_actual_file_from_ir="${TMP_DIR}/${output_base_name}_from_ir.actual_out"
+
     input_file="${sy_file%.*}.in"
     output_reference_file="${sy_file%.*}.out"
-    output_actual_file="${TMP_DIR}/${output_base_name}_sysyc_riscv64.actual_out"
 
     echo "正在处理: $(basename "$sy_file") (路径: ${relative_path_no_ext}.sy)"
 
-    # 步骤 1: 使用 sysyc 编译 .sy 到 .s
-    echo "  使用 sysyc 编译 (超时 ${SYSYC_TIMEOUT}s)..."
-    timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -S "${sy_file}" -o "${assembly_file}"
-    SYSYC_STATUS=$?
-    if [ $SYSYC_STATUS -eq 124 ]; then
-        echo -e "\e[31m错误: SysY 编译 ${sy_file} 超时\e[0m"
-        is_passed=0
-    elif [ $SYSYC_STATUS -ne 0 ]; then
-        echo -e "\e[31m错误: SysY 编译 ${sy_file} 失败，退出码: ${SYSYC_STATUS}\e[0m"
-        is_passed=0
-    fi
+    # --- 模式 1: IR 执行模式 (-eir) ---
+    if ${IR_EXECUTE_MODE}; then
+        step_failed=0
+        test_logic_passed=0
 
-    # 只有当 EXECUTE_MODE 为 true 且上一步成功时才继续
-    if ${EXECUTE_MODE} && [ "$is_passed" -eq 1 ]; then
-        # 步骤 2: 使用 riscv64-linux-gnu-gcc 编译 .s 到可执行文件
-        echo "  使用 gcc 编译 (超时 ${GCC_TIMEOUT}s)..."
-        timeout -s KILL ${GCC_TIMEOUT} "${GCC_RISCV64}" "${assembly_file}" -o "${executable_file}" -L"${LIB_DIR}" -lsysy_riscv -static
-        GCC_STATUS=$?
-        if [ $GCC_STATUS -eq 124 ]; then
-            echo -e "\e[31m错误: GCC 编译 ${assembly_file} 超时\e[0m"
-            is_passed=0
-        elif [ $GCC_STATUS -ne 0 ]; then
-            echo -e "\e[31m错误: GCC 编译 ${assembly_file} 失败，退出码: ${GCC_STATUS}\e[0m"
-            is_passed=0
+        echo "  [1/4] 使用 sysyc 编译为 IR (超时 ${SYSYC_TIMEOUT}s)..."
+        timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -s ir "${sy_file}" -o "${ir_file}" ${OPTIMIZE_FLAG}
+        SYSYC_STATUS=$?
+        if [ $SYSYC_STATUS -ne 0 ]; then
+            [ $SYSYC_STATUS -eq 124 ] && echo -e "\e[31m错误: SysY (IR) 编译超时\e[0m" || echo -e "\e[31m错误: SysY (IR) 编译失败，退出码: ${SYSYC_STATUS}\e[0m"
+            step_failed=1
         fi
-    elif ! ${EXECUTE_MODE}; then
-        echo "  跳过执行模式。仅生成汇编文件。"
-        if [ "$is_passed" -eq 1 ]; then
-             ((PASSED_CASES++))
-        else
-            FAILED_CASES_LIST+="${relative_path_no_ext}.sy\n"
-        fi
-        echo ""
-        continue
-    fi
 
-    # 步骤 3, 4, 5: 只有当编译都成功时才执行
-    if [ "$is_passed" -eq 1 ]; then
-        echo "  正在执行 (超时 ${EXEC_TIMEOUT}s)..."
-        
-        exec_cmd="${QEMU_RISCV64} \"${executable_file}\""
-        if [ -f "${input_file}" ]; then
-            exec_cmd+=" < \"${input_file}\""
-        fi
-        exec_cmd+=" > \"${output_actual_file}\""
-
-        eval "timeout -s KILL ${EXEC_TIMEOUT} ${exec_cmd}"
-        ACTUAL_RETURN_CODE=$?
-
-        if [ "$ACTUAL_RETURN_CODE" -eq 124 ]; then
-            echo -e "\e[31m  执行超时: ${sy_file} 运行超过 ${EXEC_TIMEOUT} 秒\e[0m"
-            is_passed=0
-        else
-            if [ -f "${output_reference_file}" ]; then
-                LAST_LINE_TRIMMED=$(tail -n 1 "${output_reference_file}" | tr -d '[:space:]')
-                
-                if [[ "$LAST_LINE_TRIMMED" =~ ^[-+]?[0-9]+$ ]]; then
-                    EXPECTED_RETURN_CODE="$LAST_LINE_TRIMMED"
-                    EXPECTED_STDOUT_FILE="${TMP_DIR}/${output_base_name}_sysyc_riscv64.expected_stdout"
-                    head -n -1 "${output_reference_file}" > "${EXPECTED_STDOUT_FILE}"
-
-                    if [ "$ACTUAL_RETURN_CODE" -eq "$EXPECTED_RETURN_CODE" ]; then
-                        echo -e "\e[32m  返回码测试成功: (${ACTUAL_RETURN_CODE}) 与期望值 (${EXPECTED_RETURN_CODE}) 匹配\e[0m"
-                    else
-                        echo -e "\e[31m  返回码测试失败: 期望: ${EXPECTED_RETURN_CODE}, 实际: ${ACTUAL_RETURN_CODE}\e[0m"
-                        is_passed=0
-                    fi
-
-                    if ! diff -q <(tr -d '[:space:]' < "${output_actual_file}") <(tr -d '[:space:]' < "${EXPECTED_STDOUT_FILE}") >/dev/null 2>&1; then
-                        echo -e "\e[31m  标准输出测试失败\e[0m"
-                        is_passed=0
-                        display_file_content "${EXPECTED_STDOUT_FILE}" "    \e[36m---------- 期望输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
-                        display_file_content "${output_actual_file}" "    \e[36m---------- 实际输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
-                        echo -e "    \e[36m------------------------------\e[0m"
-                    fi
-                else
-                    if [ $ACTUAL_RETURN_CODE -ne 0 ]; then
-                        echo -e "\e[33m警告: 程序以非零状态 ${ACTUAL_RETURN_CODE} 退出 (纯输出比较模式)。\e[0m"
-                    fi
-
-                    if diff -q <(tr -d '[:space:]' < "${output_actual_file}") <(tr -d '[:space:]' < "${output_reference_file}") >/dev/null 2>&1; then
-                        echo -e "\e[32m  成功: 输出与参考输出匹配\e[0m"
-                    else
-                        echo -e "\e[31m  失败: 输出不匹配\e[0m"
-                        is_passed=0
-                        display_file_content "${output_reference_file}" "    \e[36m---------- 期望输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
-                        display_file_content "${output_actual_file}" "    \e[36m---------- 实际输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
-                        echo -e "    \e[36m------------------------------\e[0m"
-                    fi
-                fi
-            else
-                echo "  无参考输出文件。程序返回码: ${ACTUAL_RETURN_CODE}"
+        if [ "$step_failed" -eq 0 ]; then
+            echo "  [2/4] 使用 llc-19 编译为汇编 (超时 ${LLC_TIMEOUT}s)..."
+            timeout -s KILL ${LLC_TIMEOUT} "${LLC_CMD}" -march=riscv64 -mcpu=generic-rv64 -mattr=+m,+a,+f,+d,+c -filetype=asm "${ir_file}" -o "${assembly_file_from_ir}"
+            LLC_STATUS=$?
+            if [ $LLC_STATUS -ne 0 ]; then
+                [ $LLC_STATUS -eq 124 ] && echo -e "\e[31m错误: llc-19 编译超时\e[0m" || echo -e "\e[31m错误: llc-19 编译失败，退出码: ${LLC_STATUS}\e[0m"
+                step_failed=1
             fi
         fi
+
+        if [ "$step_failed" -eq 0 ]; then
+            echo "  [3/4] 使用 gcc 编译 (超时 ${GCC_TIMEOUT}s)..."
+            timeout -s KILL ${GCC_TIMEOUT} "${GCC_RISCV64}" "${assembly_file_from_ir}" -o "${executable_file_from_ir}" -L"${LIB_DIR}" -lsysy_riscv -static
+            GCC_STATUS=$?
+            if [ $GCC_STATUS -ne 0 ]; then
+                [ $GCC_STATUS -eq 124 ] && echo -e "\e[31m错误: GCC 编译超时\e[0m" || echo -e "\e[31m错误: GCC 编译失败，退出码: ${GCC_STATUS}\e[0m"
+                step_failed=1
+            fi
+        fi
+
+        if [ "$step_failed" -eq 0 ]; then
+            echo "  [4/4] 正在执行 (超时 ${EXEC_TIMEOUT}s)..."
+            exec_cmd="${QEMU_RISCV64} \"${executable_file_from_ir}\""
+            [ -f "${input_file}" ] && exec_cmd+=" < \"${input_file}\""
+            exec_cmd+=" > \"${output_actual_file_from_ir}\""
+
+            eval "timeout -s KILL ${EXEC_TIMEOUT} ${exec_cmd}"
+            ACTUAL_RETURN_CODE=$?
+
+            if [ "$ACTUAL_RETURN_CODE" -eq 124 ]; then
+                echo -e "\e[31m  执行超时: 运行超过 ${EXEC_TIMEOUT} 秒\e[0m"
+            else
+                if [ -f "${output_reference_file}" ]; then
+                    LAST_LINE_TRIMMED=$(tail -n 1 "${output_reference_file}" | tr -d '[:space:]')
+                    test_logic_passed=1
+                    if [[ "$LAST_LINE_TRIMMED" =~ ^[-+]?[0-9]+$ ]]; then
+                        EXPECTED_RETURN_CODE="$LAST_LINE_TRIMMED"
+                        EXPECTED_STDOUT_FILE="${TMP_DIR}/${output_base_name}_from_ir.expected_stdout"
+                        head -n -1 "${output_reference_file}" > "${EXPECTED_STDOUT_FILE}"
+                        
+                        if [ "$ACTUAL_RETURN_CODE" -eq "$EXPECTED_RETURN_CODE" ]; then
+                            echo -e "\e[32m  返回码测试成功: (${ACTUAL_RETURN_CODE}) 与期望值 (${EXPECTED_RETURN_CODE}) 匹配\e[0m"
+                        else
+                            echo -e "\e[31m  返回码测试失败: 期望: ${EXPECTED_RETURN_CODE}, 实际: ${ACTUAL_RETURN_CODE}\e[0m"
+                            test_logic_passed=0
+                        fi
+
+                        if diff -q <(tr -d '[:space:]' < "${output_actual_file_from_ir}") <(tr -d '[:space:]' < "${EXPECTED_STDOUT_FILE}") >/dev/null 2>&1; then
+                             [ "$test_logic_passed" -eq 1 ] && echo -e "\e[32m  标准输出测试成功\e[0m"
+                        else
+                            echo -e "\e[31m  标准输出测试失败\e[0m"
+                            display_file_content "${EXPECTED_STDOUT_FILE}" "    \e[36m---------- 期望输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            display_file_content "${output_actual_file_from_ir}" "    \e[36m---------- 实际输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            test_logic_passed=0
+                        fi
+                    else
+                        if [ $ACTUAL_RETURN_CODE -ne 0 ]; then echo -e "\e[33m警告: 程序以非零状态 ${ACTUAL_RETURN_CODE} 退出 (纯输出比较模式)。\e[0m"; fi
+                        if diff -q <(tr -d '[:space:]' < "${output_actual_file_from_ir}") <(tr -d '[:space:]' < "${output_reference_file}") >/dev/null 2>&1; then
+                            echo -e "\e[32m  成功: 输出与参考输出匹配\e[0m"
+                        else
+                            echo -e "\e[31m  失败: 输出不匹配\e[0m"
+                            display_file_content "${output_reference_file}" "    \e[36m---------- 期望输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            display_file_content "${output_actual_file_from_ir}" "    \e[36m---------- 实际输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            test_logic_passed=0
+                        fi
+                    fi
+                else
+                    echo "  无参考输出文件。程序返回码: ${ACTUAL_RETURN_CODE}"
+                    test_logic_passed=1 
+                fi
+            fi
+        fi
+        [ "$step_failed" -eq 0 ] && [ "$test_logic_passed" -eq 1 ] && is_passed=1
+
+    # --- 模式 2: 直接执行模式 (-e) ---
+    elif ${EXECUTE_MODE}; then
+        step_failed=0
+        test_logic_passed=0
+
+        echo "  [1/3] 使用 sysyc 编译为汇编 (超时 ${SYSYC_TIMEOUT}s)..."
+        timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -S "${sy_file}" -o "${assembly_file_S}" ${OPTIMIZE_FLAG}
+        SYSYC_STATUS=$?
+        if [ $SYSYC_STATUS -ne 0 ]; then
+            [ $SYSYC_STATUS -eq 124 ] && echo -e "\e[31m错误: SysY (汇编) 编译超时\e[0m" || echo -e "\e[31m错误: SysY (汇编) 编译失败，退出码: ${SYSYC_STATUS}\e[0m"
+            step_failed=1
+        fi
+
+        if [ "$step_failed" -eq 0 ]; then
+            echo "  [2/3] 使用 gcc 编译 (超时 ${GCC_TIMEOUT}s)..."
+            timeout -s KILL ${GCC_TIMEOUT} "${GCC_RISCV64}" "${assembly_file_S}" -o "${executable_file_S}" -L"${LIB_DIR}" -lsysy_riscv -static
+            GCC_STATUS=$?
+            if [ $GCC_STATUS -ne 0 ]; then
+                [ $GCC_STATUS -eq 124 ] && echo -e "\e[31m错误: GCC 编译超时\e[0m" || echo -e "\e[31m错误: GCC 编译失败，退出码: ${GCC_STATUS}\e[0m"
+                step_failed=1
+            fi
+        fi
+
+        if [ "$step_failed" -eq 0 ]; then
+            echo "  [3/3] 正在执行 (超时 ${EXEC_TIMEOUT}s)..."
+            exec_cmd="${QEMU_RISCV64} \"${executable_file_S}\""
+            [ -f "${input_file}" ] && exec_cmd+=" < \"${input_file}\""
+            exec_cmd+=" > \"${output_actual_file_S}\""
+
+            eval "timeout -s KILL ${EXEC_TIMEOUT} ${exec_cmd}"
+            ACTUAL_RETURN_CODE=$?
+            
+            if [ "$ACTUAL_RETURN_CODE" -eq 124 ]; then
+                echo -e "\e[31m  执行超时: 运行超过 ${EXEC_TIMEOUT} 秒\e[0m"
+            else
+                if [ -f "${output_reference_file}" ]; then
+                    LAST_LINE_TRIMMED=$(tail -n 1 "${output_reference_file}" | tr -d '[:space:]')
+                    test_logic_passed=1
+                    if [[ "$LAST_LINE_TRIMMED" =~ ^[-+]?[0-9]+$ ]]; then
+                        EXPECTED_RETURN_CODE="$LAST_LINE_TRIMMED"
+                        EXPECTED_STDOUT_FILE="${TMP_DIR}/${output_base_name}_sysyc_S.expected_stdout"
+                        head -n -1 "${output_reference_file}" > "${EXPECTED_STDOUT_FILE}"
+                        
+                        if [ "$ACTUAL_RETURN_CODE" -eq "$EXPECTED_RETURN_CODE" ]; then
+                            echo -e "\e[32m  返回码测试成功: (${ACTUAL_RETURN_CODE}) 与期望值 (${EXPECTED_RETURN_CODE}) 匹配\e[0m"
+                        else
+                            echo -e "\e[31m  返回码测试失败: 期望: ${EXPECTED_RETURN_CODE}, 实际: ${ACTUAL_RETURN_CODE}\e[0m"
+                            test_logic_passed=0
+                        fi
+
+                        if diff -q <(tr -d '[:space:]' < "${output_actual_file_S}") <(tr -d '[:space:]' < "${EXPECTED_STDOUT_FILE}") >/dev/null 2>&1; then
+                            [ "$test_logic_passed" -eq 1 ] && echo -e "\e[32m  标准输出测试成功\e[0m"
+                        else
+                            echo -e "\e[31m  标准输出测试失败\e[0m"
+                            display_file_content "${EXPECTED_STDOUT_FILE}" "    \e[36m---------- 期望输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            display_file_content "${output_actual_file_S}" "    \e[36m---------- 实际输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            test_logic_passed=0
+                        fi
+                    else
+                        if [ $ACTUAL_RETURN_CODE -ne 0 ]; then echo -e "\e[33m警告: 程序以非零状态 ${ACTUAL_RETURN_CODE} 退出 (纯输出比较模式)。\e[0m"; fi
+                        if diff -q <(tr -d '[:space:]' < "${output_actual_file_S}") <(tr -d '[:space:]' < "${output_reference_file}") >/dev/null 2>&1; then
+                            echo -e "\e[32m  成功: 输出与参考输出匹配\e[0m"
+                        else
+                            echo -e "\e[31m  失败: 输出不匹配\e[0m"
+                            display_file_content "${output_reference_file}" "    \e[36m---------- 期望输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            display_file_content "${output_actual_file_S}" "    \e[36m---------- 实际输出 ----------\e[0m" "${MAX_OUTPUT_LINES}"
+                            test_logic_passed=0
+                        fi
+                    fi
+                else
+                    echo "  无参考输出文件。程序返回码: ${ACTUAL_RETURN_CODE}"
+                    test_logic_passed=1
+                fi
+            fi
+        fi
+        [ "$step_failed" -eq 0 ] && [ "$test_logic_passed" -eq 1 ] && is_passed=1
+
+    # --- 模式 3: 默认编译模式 ---
+    else
+        s_compile_ok=0
+        ir_compile_ok=0
+
+        echo "  [1/2] 使用 sysyc 编译为汇编 (超时 ${SYSYC_TIMEOUT}s)..."
+        timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -S "${sy_file}" -o "${assembly_file_S}" ${OPTIMIZE_FLAG}
+        SYSYC_S_STATUS=$?
+        if [ $SYSYC_S_STATUS -eq 0 ]; then
+            s_compile_ok=1
+            echo -e "      \e[32m-> ${assembly_file_S} [成功]\e[0m"
+        else
+            [ $SYSYC_S_STATUS -eq 124 ] && echo -e "      \e[31m-> [编译超时]\e[0m" || echo -e "      \e[31m-> [编译失败, 退出码: ${SYSYC_S_STATUS}]\e[0m"
+        fi
+
+        echo "  [2/2] 使用 sysyc 编译为 IR (超时 ${SYSYC_TIMEOUT}s)..."
+        timeout -s KILL ${SYSYC_TIMEOUT} "${SYSYC}" -s ir "${sy_file}" -o "${ir_file}" ${OPTIMIZE_FLAG}
+        SYSYC_IR_STATUS=$?
+        if [ $SYSYC_IR_STATUS -eq 0 ]; then
+            ir_compile_ok=1
+            echo -e "      \e[32m-> ${ir_file} [成功]\e[0m"
+        else
+            [ $SYSYC_IR_STATUS -eq 124 ] && echo -e "      \e[31m-> [编译超时]\e[0m" || echo -e "      \e[31m-> [编译失败, 退出码: ${SYSYC_IR_STATUS}]\e[0m"
+        fi
+
+        if [ "$s_compile_ok" -eq 1 ] && [ "$ir_compile_ok" -eq 1 ]; then
+            is_passed=1
+        fi
     fi
 
+    # --- 统计结果 ---
     if [ "$is_passed" -eq 1 ]; then
         ((PASSED_CASES++))
     else
+        # 确保 FAILED_CASES_LIST 的每一项都以换行符结尾
         FAILED_CASES_LIST+="${relative_path_no_ext}.sy\n"
     fi
     echo ""
 done <<< "$sy_files"
 
-echo "========================================"
-echo "测试完成"
-echo "测试通过率: [${PASSED_CASES}/${TOTAL_CASES}]"
-
-if [ -n "$FAILED_CASES_LIST" ]; then
-    echo ""
-    echo -e "\e[31m未通过的测例:\e[0m"
-    echo -e "${FAILED_CASES_LIST}"
-fi
-
-echo "========================================"
-
-if [ "$PASSED_CASES" -eq "$TOTAL_CASES" ]; then
-    exit 0
-else
-    exit 1
-fi
+# --- 修改：调用总结函数 ---
+print_summary

src/backend/RISCv64/CMakeLists.txt

@@ -12,6 +12,7 @@ add_library(riscv64_backend_lib STATIC
     Optimize/Peephole.cpp
     Optimize/PostRA_Scheduler.cpp
     Optimize/PreRA_Scheduler.cpp
+    Optimize/DivStrengthReduction.cpp
 )
 
 # 包含后端模块所需的头文件路径

src/backend/RISCv64/Handler/CalleeSavedHandler.cpp

@@ -41,7 +41,7 @@ void CalleeSavedHandler::runOnMachineFunction(MachineFunction* mfunc) {
     });
     frame_info.callee_saved_regs_to_store = sorted_regs;
 
-    // 3. [关键修正] 更新栈帧总大小。
+    // 3. 更新栈帧总大小。
     // 这是初步计算，PEI Pass 会进行最终的对齐。
     frame_info.total_size = frame_info.locals_size + 
                             frame_info.spill_size + 

src/backend/RISCv64/Handler/EliminateFrameIndices.cpp

@@ -1,7 +1,7 @@
 #include "EliminateFrameIndices.h"
 #include "RISCv64ISel.h"
 #include <cassert>
-#include <vector> // [新增] 为插入指令而包含
+#include <vector>
 
 namespace sysy {
 
@@ -33,11 +33,7 @@ void EliminateFrameIndicesPass::runOnMachineFunction(MachineFunction* mfunc) {
     Function* F = mfunc->getFunc();
     RISCv64ISel* isel = mfunc->getISel();
     
-    // 1. [已移除] 不再处理栈传递的参数
-    // 原先处理栈参数 (arg_idx >= 8) 的逻辑已被移除。
-    // 这项职责已完全转移到 PrologueEpilogueInsertionPass，以避免逻辑冲突和错误。
-    // [注释更新] -> 上述注释已过时。根据新方案，我们将在这里处理栈传递的参数，
-    // 以便在寄存器分配前就将数据流显式化，修复溢出逻辑的BUG。
+    // 在这里处理栈传递的参数，以便在寄存器分配前就将数据流显式化，修复溢出逻辑的BUG。
 
     // 2. 只为局部变量(AllocaInst)分配栈空间和计算偏移量
     // 局部变量从 s0 下方（负偏移量）开始分配，紧接着为 ra 和 s0 预留的16字节之后
@@ -50,9 +46,13 @@ void EliminateFrameIndicesPass::runOnMachineFunction(MachineFunction* mfunc) {
                     Type* allocated_type = alloca->getType()->as<PointerType>()->getBaseType();
                     int size = getTypeSizeInBytes(allocated_type);
                     
-                    // RISC-V要求栈地址8字节对齐
-                    size = (size + 7) & ~7;
-                    if (size == 0) size = 8; // 至少分配8字节
+                    // 优化栈帧大小：对于大数组使用4字节对齐，小对象使用8字节对齐
+                    if (size >= 256) {  // 大数组优化
+                        size = (size + 3) & ~3;  // 4字节对齐
+                    } else {
+                        size = (size + 7) & ~7;  // 8字节对齐
+                    }
+                    if (size == 0) size = 4; // 最小4字节
 
                     local_var_offset += size;
                     unsigned alloca_vreg = isel->getVReg(alloca);
@@ -65,13 +65,17 @@ void EliminateFrameIndicesPass::runOnMachineFunction(MachineFunction* mfunc) {
     
     // 记录仅由AllocaInst分配的局部变量的总大小
     frame_info.locals_size = local_var_offset - 16;
+    // 记录局部变量区域分配结束的最终偏移量
+    frame_info.locals_end_offset = -local_var_offset;
 
-    // 3. [核心修改] 在函数入口为所有栈传递的参数插入load指令
+    // 在函数入口为所有栈传递的参数插入load指令
     // 这个步骤至关重要：它在寄存器分配之前，为这些参数的vreg创建了明确的“定义(def)”指令。
     // 这解决了在高寄存器压力下，当这些vreg被溢出时，`rewriteProgram`找不到其定义点而崩溃的问题。
     if (F && isel && !mfunc->getBlocks().empty()) {
         MachineBasicBlock* entry_block = mfunc->getBlocks().front().get();
         std::vector<std::unique_ptr<MachineInstr>> arg_load_instrs;
+        
+        // 步骤 3.1: 生成所有加载栈参数的指令，暂存起来
         int arg_idx = 0;
         for (Argument* arg : F->getArguments()) {
             // 根据ABI，前8个整型/指针参数通过寄存器传递，这里只处理超出部分。
@@ -104,17 +108,50 @@ void EliminateFrameIndicesPass::runOnMachineFunction(MachineFunction* mfunc) {
             arg_idx++;
         }
         
-        // 将所有新创建的参数加载指令一次性插入到入口块的起始位置
-        auto& entry_instrs = entry_block->getInstructions();
-        entry_instrs.insert(entry_instrs.begin(),
-                            std::make_move_iterator(arg_load_instrs.begin()),
-                            std::make_move_iterator(arg_load_instrs.end()));
+        //仅当有需要加载的栈参数时，才执行插入逻辑
+        if (!arg_load_instrs.empty()) {
+            auto& entry_instrs = entry_block->getInstructions();
+            auto insertion_point = entry_instrs.begin(); // 默认插入点为块的开头
+            auto last_arg_save_it = entry_instrs.end();
+
+            // 步骤 3.2: 寻找一个安全的插入点。
+            // 遍历入口块的指令，找到最后一条保存“寄存器传递参数”的伪指令。
+            // 这样可以确保我们在所有 a0-a7 参数被保存之后，才执行可能覆盖它们的加载指令。
+            for (auto it = entry_instrs.begin(); it != entry_instrs.end(); ++it) {
+                MachineInstr* instr = it->get();
+                // 寻找代表保存参数到栈的伪指令
+                if (instr->getOpcode() == RVOpcodes::FRAME_STORE_W ||
+                    instr->getOpcode() == RVOpcodes::FRAME_STORE_D ||
+                    instr->getOpcode() == RVOpcodes::FRAME_STORE_F) {
+                    
+                    // 检查被保存的值是否是寄存器参数 (arg_no < 8)
+                    auto& operands = instr->getOperands();
+                    if (operands.empty() || operands[0]->getKind() != MachineOperand::KIND_REG) continue;
+                    
+                    unsigned src_vreg = static_cast<RegOperand*>(operands[0].get())->getVRegNum();
+                    Value* ir_value = isel->getVRegValueMap().count(src_vreg) ? isel->getVRegValueMap().at(src_vreg) : nullptr;
+                    
+                    if (auto ir_arg = dynamic_cast<Argument*>(ir_value)) {
+                        if (ir_arg->getIndex() < 8) {
+                            last_arg_save_it = it; // 找到了一个保存寄存器参数的指令，更新位置
+                        }
+                    }
+                }
+            }
+
+            // 如果找到了这样的保存指令，我们的插入点就在它之后
+            if (last_arg_save_it != entry_instrs.end()) {
+                insertion_point = std::next(last_arg_save_it);
+            }
+
+            // 步骤 3.3: 在计算出的安全位置，一次性插入所有新创建的参数加载指令
+            entry_instrs.insert(insertion_point,
+                                std::make_move_iterator(arg_load_instrs.begin()),
+                                std::make_move_iterator(arg_load_instrs.end()));
+        }
     }
 
-
     // 4. 遍历所有机器指令，将访问局部变量的伪指令展开为真实指令
-    // 由于处理参数的逻辑已移除，这里的展开现在只针对局部变量，因此是正确的。
-    // [注释更新] -> 上述注释已过时。此部分逻辑保持不变，它正确地处理了局部变量。
     for (auto& mbb : mfunc->getBlocks()) {
         std::vector<std::unique_ptr<MachineInstr>> new_instructions;
         for (auto& instr_ptr : mbb->getInstructions()) {

src/backend/RISCv64/Handler/PrologueEpilogueInsertion.cpp

@@ -27,8 +27,7 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
         );
     }
     
-    // 2. [新增] 确定需要保存的被调用者保存寄存器 (callee-saved)
-    // 这部分逻辑从 CalleeSavedHandler Pass 移入，以集中管理序言生成
+    // 2. 确定需要保存的被调用者保存寄存器 (callee-saved)
     auto& vreg_to_preg_map = frame_info.vreg_to_preg_map;
     std::set<PhysicalReg> used_callee_saved_regs_set;
     const auto& callee_saved_int = getCalleeSavedIntRegs();
@@ -36,38 +35,44 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
 
     for (const auto& pair : vreg_to_preg_map) {
         PhysicalReg preg = pair.second;
-        // 检查是否在整数或浮点 callee-saved 集合中
-        // 注意：s0作为帧指针，由序言/尾声逻辑特殊处理，不在此处保存
         bool is_int_cs = std::find(callee_saved_int.begin(), callee_saved_int.end(), preg) != callee_saved_int.end();
         bool is_fp_cs = std::find(callee_saved_fp.begin(), callee_saved_fp.end(), preg) != callee_saved_fp.end();
         if ((is_int_cs && preg != PhysicalReg::S0) || is_fp_cs) {
             used_callee_saved_regs_set.insert(preg);
         }
     }
-    // 为了确定性排序，并存入 frame_info 供尾声使用
     frame_info.callee_saved_regs_to_store.assign(
         used_callee_saved_regs_set.begin(), used_callee_saved_regs_set.end()
     );
     std::sort(frame_info.callee_saved_regs_to_store.begin(), frame_info.callee_saved_regs_to_store.end());
-    frame_info.callee_saved_size = frame_info.callee_saved_regs_to_store.size() * 8; // 每个寄存器8字节
+    frame_info.callee_saved_size = frame_info.callee_saved_regs_to_store.size() * 8;
 
-    // 3. 计算最终的栈帧总大小
+    // 3. 计算最终的栈帧总大小，包含栈溢出保护
     int total_stack_size = frame_info.locals_size + 
                            frame_info.spill_size + 
                            frame_info.callee_saved_size + 
-                           16; // 为 ra 和 s0 固定的16字节
+                           16;
     
-    int aligned_stack_size = (total_stack_size + 15) & ~15; // 16字节对齐
+    // 栈溢出保护：增加最大栈帧大小以容纳大型数组
+    const int MAX_STACK_FRAME_SIZE = 8192; // 8KB to handle large arrays like 256*4*2 = 2048 bytes
+    if (total_stack_size > MAX_STACK_FRAME_SIZE) {
+        // 如果仍然超过限制，尝试优化对齐方式
+        std::cerr << "Warning: Stack frame size " << total_stack_size 
+                  << " exceeds recommended limit " << MAX_STACK_FRAME_SIZE << " for function " 
+                  << mfunc->getName() << std::endl;
+    }
+    
+    // 优化：减少对齐开销，使用16字节对齐而非更大的对齐
+    int aligned_stack_size = (total_stack_size + 15) & ~15;
     frame_info.total_size = aligned_stack_size;
 
-    // 只有在需要分配栈空间时才生成序言和尾声
     if (aligned_stack_size > 0) {
         // --- 4. 插入完整的序言 ---
         MachineBasicBlock* entry_block = mfunc->getBlocks().front().get();
         auto& entry_instrs = entry_block->getInstructions();
         std::vector<std::unique_ptr<MachineInstr>> prologue_instrs;
 
-        // 4.1. 分配栈帧: addi sp, sp, -aligned_stack_size
+        // 4.1. 分配栈帧
         auto alloc_stack = std::make_unique<MachineInstr>(RVOpcodes::ADDI);
         alloc_stack->addOperand(std::make_unique<RegOperand>(PhysicalReg::SP));
         alloc_stack->addOperand(std::make_unique<RegOperand>(PhysicalReg::SP));
@@ -75,7 +80,6 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
         prologue_instrs.push_back(std::move(alloc_stack));
 
         // 4.2. 保存 ra 和 s0
-        // sd ra, (aligned_stack_size - 8)(sp)
         auto save_ra = std::make_unique<MachineInstr>(RVOpcodes::SD);
         save_ra->addOperand(std::make_unique<RegOperand>(PhysicalReg::RA));
         save_ra->addOperand(std::make_unique<MemOperand>(
@@ -83,7 +87,6 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
             std::make_unique<ImmOperand>(aligned_stack_size - 8)
         ));
         prologue_instrs.push_back(std::move(save_ra));
-        // sd s0, (aligned_stack_size - 16)(sp)
         auto save_fp = std::make_unique<MachineInstr>(RVOpcodes::SD);
         save_fp->addOperand(std::make_unique<RegOperand>(PhysicalReg::S0));
         save_fp->addOperand(std::make_unique<MemOperand>(
@@ -92,60 +95,55 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
         ));
         prologue_instrs.push_back(std::move(save_fp));
         
-        // 4.3. 设置新的帧指针 s0: addi s0, sp, aligned_stack_size
+        // 4.3. 设置新的帧指针 s0
         auto set_fp = std::make_unique<MachineInstr>(RVOpcodes::ADDI);
         set_fp->addOperand(std::make_unique<RegOperand>(PhysicalReg::S0));
         set_fp->addOperand(std::make_unique<RegOperand>(PhysicalReg::SP));
         set_fp->addOperand(std::make_unique<ImmOperand>(aligned_stack_size));
         prologue_instrs.push_back(std::move(set_fp));
         
-        // 4.4. [新增] 保存所有使用到的被调用者保存寄存器
-        // 它们保存在 s0 下方，紧接着 ra/s0 的位置
-        int callee_saved_offset = -16;
+        // 4.4. 保存所有使用到的被调用者保存寄存器
+        int next_available_offset = -(16 + frame_info.locals_size + frame_info.spill_size);
         for (const auto& reg : frame_info.callee_saved_regs_to_store) {
-            callee_saved_offset -= 8;
-            RVOpcodes store_op = (reg >= PhysicalReg::F0 && reg <= PhysicalReg::F31) ? RVOpcodes::FSD : RVOpcodes::SD;
+            // 改为“先更新，后使用”逻辑
+            next_available_offset -= 8; // 先为当前寄存器分配下一个可用槽位
+            RVOpcodes store_op = isFPR(reg) ? RVOpcodes::FSD : RVOpcodes::SD;
             auto save_cs_reg = std::make_unique<MachineInstr>(store_op);
             save_cs_reg->addOperand(std::make_unique<RegOperand>(reg));
             save_cs_reg->addOperand(std::make_unique<MemOperand>(
                 std::make_unique<RegOperand>(PhysicalReg::S0),
-                std::make_unique<ImmOperand>(callee_saved_offset)
+                std::make_unique<ImmOperand>(next_available_offset) // 使用新计算出的正确偏移
             ));
             prologue_instrs.push_back(std::move(save_cs_reg));
+            // 不再需要在循环末尾递减
         }
 
-        // 4.5. [核心修改] 加载栈传递参数的逻辑已从此移除
-        // 这项工作已经前移至 `EliminateFrameIndicesPass` 中完成，
-        // 以确保在寄存器分配前就将相关虚拟寄存器定义，从而修复溢出逻辑的bug。
-        
-        // 4.6. 将所有生成的序言指令一次性插入到函数入口
+        // 4.5. 将所有生成的序言指令一次性插入到函数入口
         entry_instrs.insert(entry_instrs.begin(), 
                             std::make_move_iterator(prologue_instrs.begin()),
                             std::make_move_iterator(prologue_instrs.end()));
 
         // --- 5. 插入完整的尾声 ---
         for (auto& mbb : mfunc->getBlocks()) {
-            // [修正] 使用前向迭代器查找RET指令，以确保在正确的位置（RET之前）插入尾声。
             for (auto it = mbb->getInstructions().begin(); it != mbb->getInstructions().end(); ++it) {
                 if ((*it)->getOpcode() == RVOpcodes::RET) {
                     std::vector<std::unique_ptr<MachineInstr>> epilogue_instrs;
                     
-                    // 5.1. [新增] 恢复被调用者保存寄存器
-                    callee_saved_offset = -16;
+                    // 5.1. 恢复被调用者保存寄存器
+                    int next_available_offset_restore = -(16 + frame_info.locals_size + frame_info.spill_size);
                     for (const auto& reg : frame_info.callee_saved_regs_to_store) {
-                        callee_saved_offset -= 8;
-                        RVOpcodes load_op = (reg >= PhysicalReg::F0 && reg <= PhysicalReg::F31) ? RVOpcodes::FLD : RVOpcodes::LD;
+                        next_available_offset_restore -= 8; // 为下一个寄存器准备偏移
+                        RVOpcodes load_op = isFPR(reg) ? RVOpcodes::FLD : RVOpcodes::LD;
                         auto restore_cs_reg = std::make_unique<MachineInstr>(load_op);
                         restore_cs_reg->addOperand(std::make_unique<RegOperand>(reg));
                         restore_cs_reg->addOperand(std::make_unique<MemOperand>(
                             std::make_unique<RegOperand>(PhysicalReg::S0),
-                            std::make_unique<ImmOperand>(callee_saved_offset)
+                            std::make_unique<ImmOperand>(next_available_offset_restore) // 使用当前偏移
                         ));
                         epilogue_instrs.push_back(std::move(restore_cs_reg));
                     }
 
-                    // 5.2. 恢复 ra 和 s0 (注意基址现在是sp)
-                    // ld ra, (aligned_stack_size - 8)(sp)
+                    // 5.2. 恢复 ra 和 s0
                     auto restore_ra = std::make_unique<MachineInstr>(RVOpcodes::LD);
                     restore_ra->addOperand(std::make_unique<RegOperand>(PhysicalReg::RA));
                     restore_ra->addOperand(std::make_unique<MemOperand>(
@@ -153,7 +151,6 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
                         std::make_unique<ImmOperand>(aligned_stack_size - 8)
                     ));
                     epilogue_instrs.push_back(std::move(restore_ra));
-                    // ld s0, (aligned_stack_size - 16)(sp)
                     auto restore_fp = std::make_unique<MachineInstr>(RVOpcodes::LD);
                     restore_fp->addOperand(std::make_unique<RegOperand>(PhysicalReg::S0));
                     restore_fp->addOperand(std::make_unique<MemOperand>(
@@ -162,7 +159,7 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
                     ));
                     epilogue_instrs.push_back(std::move(restore_fp));
 
-                    // 5.3. 释放栈帧: addi sp, sp, aligned_stack_size
+                    // 5.3. 释放栈帧
                     auto dealloc_stack = std::make_unique<MachineInstr>(RVOpcodes::ADDI);
                     dealloc_stack->addOperand(std::make_unique<RegOperand>(PhysicalReg::SP));
                     dealloc_stack->addOperand(std::make_unique<RegOperand>(PhysicalReg::SP));
@@ -174,7 +171,6 @@ void PrologueEpilogueInsertionPass::runOnMachineFunction(MachineFunction* mfunc)
                                                   std::make_move_iterator(epilogue_instrs.begin()),
                                                   std::make_move_iterator(epilogue_instrs.end()));
                     
-                    // 一个基本块通常只有一个终止指令，处理完就可以跳到下一个块
                     goto next_block;
                 }
             }

src/backend/RISCv64/Optimize/DivStrengthReduction.cpp

@@ -0,0 +1,282 @@
+#include "DivStrengthReduction.h"
+#include <cmath>
+#include <cstdint>
+
+namespace sysy {
+
+char DivStrengthReduction::ID = 0;
+
+bool DivStrengthReduction::runOnFunction(Function *F, AnalysisManager& AM) {
+    // This pass works on MachineFunction level, not IR level
+    return false;
+}
+
+void DivStrengthReduction::runOnMachineFunction(MachineFunction *mfunc) {
+    if (!mfunc)
+        return;
+
+    bool debug = false; // Set to true for debugging
+    if (debug)
+        std::cout << "Running DivStrengthReduction optimization..." << std::endl;
+
+    int next_temp_reg = 1000;
+    auto createTempReg = [&]() -> int {
+        return next_temp_reg++;
+    };
+
+    struct MagicInfo {
+        int64_t magic;
+        int shift;
+    };
+    
+    auto computeMagic = [](int64_t d, bool is_32bit) -> MagicInfo {
+        int word_size = is_32bit ? 32 : 64;
+        uint64_t ad = std::abs(d);
+        
+        if (ad == 0) return {0, 0};
+        
+        int l = std::floor(std::log2(ad));
+        if ((ad & (ad - 1)) == 0) { // power of 2
+             l = 0; // special case for power of 2, shift will be calculated differently
+        }
+
+        __int128_t one = 1;
+        __int128_t num;
+        int total_shift;
+
+        if (is_32bit) {
+            total_shift = 31 + l;
+            num = one << total_shift;
+        } else {
+            total_shift = 63 + l;
+            num = one << total_shift;
+        }
+        
+        __int128_t den = ad;
+        int64_t magic = (num / den) + 1;
+        
+        return {magic, total_shift};
+    };
+
+    auto isPowerOfTwo = [](int64_t n) -> bool {
+        return n > 0 && (n & (n - 1)) == 0;
+    };
+
+    auto getPowerOfTwoExponent = [](int64_t n) -> int {
+        if (n <= 0 || (n & (n - 1)) != 0) return -1;
+        int shift = 0;
+        while (n > 1) {
+            n >>= 1;
+            shift++;
+        }
+        return shift;
+    };
+
+    struct InstructionReplacement {
+        size_t index;
+        size_t count_to_erase;
+        std::vector<std::unique_ptr<MachineInstr>> newInstrs;
+    };
+    
+    for (auto &mbb_uptr : mfunc->getBlocks()) {
+        auto &mbb = *mbb_uptr;
+        auto &instrs = mbb.getInstructions();
+        std::vector<InstructionReplacement> replacements;
+        
+        for (size_t i = 0; i < instrs.size(); ++i) {
+            auto *instr = instrs[i].get();
+            
+            bool is_32bit = (instr->getOpcode() == RVOpcodes::DIVW);
+            
+            if (instr->getOpcode() != RVOpcodes::DIV && !is_32bit) {
+                continue;
+            }
+            
+            if (instr->getOperands().size() != 3) {
+                continue;
+            }
+            
+            auto *dst_op = instr->getOperands()[0].get();
+            auto *src1_op = instr->getOperands()[1].get();
+            auto *src2_op = instr->getOperands()[2].get();
+
+            int64_t divisor = 0;
+            bool const_divisor_found = false;
+            size_t instructions_to_replace = 1;
+
+            if (src2_op->getKind() == MachineOperand::KIND_IMM) {
+                divisor = static_cast<ImmOperand *>(src2_op)->getValue();
+                const_divisor_found = true;
+            } else if (src2_op->getKind() == MachineOperand::KIND_REG) {
+                if (i > 0) {
+                    auto *prev_instr = instrs[i - 1].get();
+                    if (prev_instr->getOpcode() == RVOpcodes::LI && prev_instr->getOperands().size() == 2) {
+                        auto *li_dst_op = prev_instr->getOperands()[0].get();
+                        auto *li_imm_op = prev_instr->getOperands()[1].get();
+                        if (li_dst_op->getKind() == MachineOperand::KIND_REG && li_imm_op->getKind() == MachineOperand::KIND_IMM) {
+                            auto *div_reg_op = static_cast<RegOperand *>(src2_op);
+                            auto *li_dst_reg_op = static_cast<RegOperand *>(li_dst_op);
+                            if (div_reg_op->isVirtual() && li_dst_reg_op->isVirtual() &&
+                                div_reg_op->getVRegNum() == li_dst_reg_op->getVRegNum()) {
+                                divisor = static_cast<ImmOperand *>(li_imm_op)->getValue();
+                                const_divisor_found = true;
+                                instructions_to_replace = 2;
+                            }
+                        }
+                    }
+                }
+            }
+
+            if (!const_divisor_found) {
+                continue;
+            }
+            
+            auto *dst_reg = static_cast<RegOperand *>(dst_op);
+            auto *src1_reg = static_cast<RegOperand *>(src1_op);
+            
+            if (divisor == 0) continue;
+            
+            std::vector<std::unique_ptr<MachineInstr>> newInstrs;
+            
+            if (divisor == 1) {
+                auto moveInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::ADDW : RVOpcodes::ADD);
+                moveInstr->addOperand(std::make_unique<RegOperand>(*dst_reg));
+                moveInstr->addOperand(std::make_unique<RegOperand>(*src1_reg));
+                moveInstr->addOperand(std::make_unique<RegOperand>(PhysicalReg::ZERO));
+                newInstrs.push_back(std::move(moveInstr));
+            }
+            else if (divisor == -1) {
+                auto negInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SUBW : RVOpcodes::SUB);
+                negInstr->addOperand(std::make_unique<RegOperand>(*dst_reg));
+                negInstr->addOperand(std::make_unique<RegOperand>(PhysicalReg::ZERO));
+                negInstr->addOperand(std::make_unique<RegOperand>(*src1_reg));
+                newInstrs.push_back(std::move(negInstr));
+            }
+            else if (isPowerOfTwo(std::abs(divisor))) {
+                int shift = getPowerOfTwoExponent(std::abs(divisor));
+                int temp_reg = createTempReg();
+                
+                auto sraSignInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SRAIW : RVOpcodes::SRAI);
+                sraSignInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                sraSignInstr->addOperand(std::make_unique<RegOperand>(*src1_reg));
+                sraSignInstr->addOperand(std::make_unique<ImmOperand>(is_32bit ? 31 : 63));
+                newInstrs.push_back(std::move(sraSignInstr));
+                
+                auto srlInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SRLIW : RVOpcodes::SRLI);
+                srlInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                srlInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                srlInstr->addOperand(std::make_unique<ImmOperand>((is_32bit ? 32 : 64) - shift));
+                newInstrs.push_back(std::move(srlInstr));
+                
+                auto addInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::ADDW : RVOpcodes::ADD);
+                addInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                addInstr->addOperand(std::make_unique<RegOperand>(*src1_reg));
+                addInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                newInstrs.push_back(std::move(addInstr));
+                
+                auto sraInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SRAIW : RVOpcodes::SRAI);
+                sraInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                sraInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                sraInstr->addOperand(std::make_unique<ImmOperand>(shift));
+                newInstrs.push_back(std::move(sraInstr));
+
+                if (divisor < 0) {
+                    auto negInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SUBW : RVOpcodes::SUB);
+                    negInstr->addOperand(std::make_unique<RegOperand>(*dst_reg));
+                    negInstr->addOperand(std::make_unique<RegOperand>(PhysicalReg::ZERO));
+                    negInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    newInstrs.push_back(std::move(negInstr));
+                } else {
+                    auto moveInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::ADDW : RVOpcodes::ADD);
+                    moveInstr->addOperand(std::make_unique<RegOperand>(*dst_reg));
+                    moveInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    moveInstr->addOperand(std::make_unique<RegOperand>(PhysicalReg::ZERO));
+                    newInstrs.push_back(std::move(moveInstr));
+                }
+            }
+            else {
+                auto magic_info = computeMagic(divisor, is_32bit);
+                int magic_reg = createTempReg();
+                int temp_reg = createTempReg();
+
+                auto loadInstr = std::make_unique<MachineInstr>(RVOpcodes::LI);
+                loadInstr->addOperand(std::make_unique<RegOperand>(magic_reg));
+                loadInstr->addOperand(std::make_unique<ImmOperand>(magic_info.magic));
+                newInstrs.push_back(std::move(loadInstr));
+
+                if (is_32bit) {
+                    auto mulInstr = std::make_unique<MachineInstr>(RVOpcodes::MUL);
+                    mulInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    mulInstr->addOperand(std::make_unique<RegOperand>(*src1_reg));
+                    mulInstr->addOperand(std::make_unique<RegOperand>(magic_reg));
+                    newInstrs.push_back(std::move(mulInstr));
+
+                    auto sraInstr = std::make_unique<MachineInstr>(RVOpcodes::SRAI);
+                    sraInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    sraInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    sraInstr->addOperand(std::make_unique<ImmOperand>(magic_info.shift));
+                    newInstrs.push_back(std::move(sraInstr));
+                } else {
+                    auto mulhInstr = std::make_unique<MachineInstr>(RVOpcodes::MULH);
+                    mulhInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    mulhInstr->addOperand(std::make_unique<RegOperand>(*src1_reg));
+                    mulhInstr->addOperand(std::make_unique<RegOperand>(magic_reg));
+                    newInstrs.push_back(std::move(mulhInstr));
+                    
+                    int post_shift = magic_info.shift - 63;
+                    if (post_shift > 0) {
+                        auto sraInstr = std::make_unique<MachineInstr>(RVOpcodes::SRAI);
+                        sraInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                        sraInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                        sraInstr->addOperand(std::make_unique<ImmOperand>(post_shift));
+                        newInstrs.push_back(std::move(sraInstr));
+                    }
+                }
+                
+                int sign_reg = createTempReg();
+                auto sraSignInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SRAIW : RVOpcodes::SRAI);
+                sraSignInstr->addOperand(std::make_unique<RegOperand>(sign_reg));
+                sraSignInstr->addOperand(std::make_unique<RegOperand>(*src1_reg));
+                sraSignInstr->addOperand(std::make_unique<ImmOperand>(is_32bit ? 31 : 63));
+                newInstrs.push_back(std::move(sraSignInstr));
+
+                auto subInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SUBW : RVOpcodes::SUB);
+                subInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                subInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                subInstr->addOperand(std::make_unique<RegOperand>(sign_reg));
+                newInstrs.push_back(std::move(subInstr));
+
+                if (divisor < 0) {
+                    auto negInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::SUBW : RVOpcodes::SUB);
+                    negInstr->addOperand(std::make_unique<RegOperand>(*dst_reg));
+                    negInstr->addOperand(std::make_unique<RegOperand>(PhysicalReg::ZERO));
+                    negInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    newInstrs.push_back(std::move(negInstr));
+                } else {
+                    auto moveInstr = std::make_unique<MachineInstr>(is_32bit ? RVOpcodes::ADDW : RVOpcodes::ADD);
+                    moveInstr->addOperand(std::make_unique<RegOperand>(*dst_reg));
+                    moveInstr->addOperand(std::make_unique<RegOperand>(temp_reg));
+                    moveInstr->addOperand(std::make_unique<RegOperand>(PhysicalReg::ZERO));
+                    newInstrs.push_back(std::move(moveInstr));
+                }
+            }
+            
+            if (!newInstrs.empty()) {
+                size_t start_index = i;
+                if (instructions_to_replace == 2) {
+                    start_index = i - 1;
+                }
+                replacements.push_back({start_index, instructions_to_replace, std::move(newInstrs)});
+            }
+        }
+        
+        for (auto it = replacements.rbegin(); it != replacements.rend(); ++it) {
+            instrs.erase(instrs.begin() + it->index, instrs.begin() + it->index + it->count_to_erase);
+            instrs.insert(instrs.begin() + it->index, 
+                         std::make_move_iterator(it->newInstrs.begin()),
+                         std::make_move_iterator(it->newInstrs.end()));
+        }
+    }
+}
+
+} // namespace sysy

src/backend/RISCv64/Optimize/Peephole.cpp

@@ -634,6 +634,22 @@ void PeepholeOptimizer::runOnMachineFunction(MachineFunction *mfunc) {
           }
         }
       }
+      // 8. 消除无用移动指令: mv a, a -> (删除)
+      else if (mi1->getOpcode() == RVOpcodes::MV &&
+               mi1->getOperands().size() == 2) {
+        if (mi1->getOperands()[0]->getKind() == MachineOperand::KIND_REG &&
+            mi1->getOperands()[1]->getKind() == MachineOperand::KIND_REG) {
+          auto *dst = static_cast<RegOperand *>(mi1->getOperands()[0].get());
+          auto *src = static_cast<RegOperand *>(mi1->getOperands()[1].get());
+          
+          // 检查源和目标寄存器是否相同
+          if (areRegsEqual(dst, src)) {
+            // 删除这条无用指令
+            instrs.erase(instrs.begin() + i);
+            changed = true;
+          }
+        }
+      }
 
       // 根据是否发生变化调整遍历索引
       if (!changed) {

src/backend/RISCv64/RISCv64AsmPrinter.cpp

@@ -1,7 +1,8 @@
 #include "RISCv64AsmPrinter.h"
 #include "RISCv64ISel.h"
 #include <stdexcept>
-
+#include <sstream>
+#include <iostream>
 namespace sysy {
 
 // 检查是否为内存加载/存储指令，以处理特殊的打印格式
@@ -60,7 +61,7 @@ void RISCv64AsmPrinter::printInstruction(MachineInstr* instr, bool debug) {
         case RVOpcodes::ADD:   *OS << "add ";   break; case RVOpcodes::ADDI:  *OS << "addi ";  break;
         case RVOpcodes::ADDW:  *OS << "addw ";  break; case RVOpcodes::ADDIW: *OS << "addiw "; break;
         case RVOpcodes::SUB:   *OS << "sub ";   break; case RVOpcodes::SUBW:  *OS << "subw ";  break;
-        case RVOpcodes::MUL:   *OS << "mul ";   break; case RVOpcodes::MULW:  *OS << "mulw ";  break;
+        case RVOpcodes::MUL:   *OS << "mul ";   break; case RVOpcodes::MULW:  *OS << "mulw ";  break; case RVOpcodes::MULH:  *OS << "mulh ";  break;
         case RVOpcodes::DIV:   *OS << "div ";   break; case RVOpcodes::DIVW:  *OS << "divw ";  break;
         case RVOpcodes::REM:   *OS << "rem ";   break; case RVOpcodes::REMW:  *OS << "remw ";  break;
         case RVOpcodes::XOR:   *OS << "xor ";   break; case RVOpcodes::XORI:  *OS << "xori ";  break;
@@ -104,7 +105,7 @@ void RISCv64AsmPrinter::printInstruction(MachineInstr* instr, bool debug) {
         case RVOpcodes::FMV_S:    *OS << "fmv.s ";    break;
         case RVOpcodes::FMV_W_X:  *OS << "fmv.w.x ";  break;
         case RVOpcodes::FMV_X_W:  *OS << "fmv.x.w ";  break;
-        case RVOpcodes::CALL: { // [核心修改] 为CALL指令添加特殊处理逻辑
+        case RVOpcodes::CALL: { // 为CALL指令添加特殊处理逻辑
             *OS << "call ";
             // 遍历所有操作数，只寻找并打印函数名标签
             for (const auto& op : instr->getOperands()) {
@@ -236,4 +237,30 @@ std::string RISCv64AsmPrinter::regToString(PhysicalReg reg) {
     }
 }
 
+std::string RISCv64AsmPrinter::formatInstr(const MachineInstr* instr) {
+    if (!instr) return "(null instr)";
+    
+    // 使用 stringstream 作为临时的输出目标
+    std::stringstream ss;
+    
+    // 关键: 临时将类成员 'OS' 指向我们的 stringstream
+    std::ostream* old_os = this->OS;
+    this->OS = &ss;
+    
+    // 修正: 调用正确的内部打印函数 printMachineInstr
+    printInstruction(const_cast<MachineInstr*>(instr), false);
+    
+    // 恢复旧的 ostream 指针
+    this->OS = old_os;
+    
+    // 获取stringstream的内容并做一些清理
+    std::string result = ss.str();
+    size_t endpos = result.find_last_not_of(" \t\n\r");
+    if (std::string::npos != endpos) {
+        result = result.substr(0, endpos + 1);
+    }
+    
+    return result;
+}
+
 } // namespace sysy

src/backend/RISCv64/RISCv64Backend.cpp

@@ -73,7 +73,7 @@ std::string RISCv64CodeGen::module_gen() {
     for (const auto& global_ptr : module->getGlobals()) {
         GlobalValue* global = global_ptr.get();
         
-        // [核心修改] 使用更健壮的逻辑来判断是否为大型零初始化数组
+        // 使用更健壮的逻辑来判断是否为大型零初始化数组
         bool is_all_zeros = true;
         const auto& init_values = global->getInitValues();
         
@@ -139,7 +139,29 @@ std::string RISCv64CodeGen::module_gen() {
             ss << ".type " << global->getName() << ", @object\n";
             ss << ".size " << global->getName() << ", " << total_size << "\n";
             ss << global->getName() << ":\n";
-            printInitializer(ss, global->getInitValues());
+            bool is_all_zeros = true;
+            const auto& init_values = global->getInitValues();
+            if (init_values.getValues().empty()) {
+                is_all_zeros = true;
+            } else {
+                for (auto val : init_values.getValues()) {
+                    if (auto const_val = dynamic_cast<ConstantValue*>(val)) {
+                        if (!const_val->isZero()) {
+                            is_all_zeros = false;
+                            break;
+                        }
+                    } else {
+                        is_all_zeros = false;
+                        break;
+                    }
+                }
+            }
+            if (is_all_zeros) {
+                ss << "    .zero " << total_size << "\n";
+            } else {
+                // 对于有非零初始值的变量，保持原有的打印逻辑。
+                printInitializer(ss, global->getInitValues());
+            }
         }
 
         // b. 处理全局常量 (ConstantVariable)
@@ -171,222 +193,90 @@ std::string RISCv64CodeGen::module_gen() {
 }
 
 std::string RISCv64CodeGen::function_gen(Function* func) {
+    // === 完整的后端处理流水线 ===
+
+    // 阶段 1: 指令选择 (sysy::IR -> LLIR with virtual registers)
+    DEBUG = 0;
+    DEEPDEBUG = 0;
+
+    RISCv64ISel isel;
+    std::unique_ptr<MachineFunction> mfunc = isel.runOnFunction(func);
+
+    // 第一次调试打印输出
+    std::stringstream ss_after_isel;
+    RISCv64AsmPrinter printer_isel(mfunc.get());
+    printer_isel.run(ss_after_isel, true);
     if (DEBUG) {
-        // === 完整的后端处理流水线 ===
-
-        // 阶段 1: 指令选择 (sysy::IR -> LLIR with virtual registers)
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-
-        RISCv64ISel isel;
-        std::unique_ptr<MachineFunction> mfunc = isel.runOnFunction(func);
-
-        // 第一次调试打印输出
-        std::stringstream ss_after_isel;
-        RISCv64AsmPrinter printer_isel(mfunc.get());
-        printer_isel.run(ss_after_isel, true);
-        // if (DEBUG) {
-        //     std::cout << ss_after_isel.str();
-        // }
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-        DEBUG = 1;
-        DEEPDEBUG = 1;
-        if (DEBUG) {
-            std::cerr << "====== Intermediate Representation after Instruction Selection ======\n" 
-            << ss_after_isel.str();
-        }
-        // DEBUG = 0;
-        // DEEPDEBUG = 0;
-        // [新增] 阶段 2: 消除帧索引 (展开伪指令，计算局部变量偏移)
-        // 这个Pass必须在寄存器分配之前运行
-        EliminateFrameIndicesPass efi_pass;
-        efi_pass.runOnMachineFunction(mfunc.get());
-
-        if (DEBUG) {
-            std::cerr << "====== stack info after eliminate frame indices  ======\n";
-            mfunc->dumpStackFrameInfo(std::cerr);
-            // std::stringstream ss_after_eli;
-            // printer_isel.run(ss_after_eli, true);
-            // std::cerr << "====== LLIR after eliminate frame indices ======\n" 
-            // << ss_after_eli.str();
-        }
-
-        // // 阶段 2: 指令调度 (Instruction Scheduling)
-        // PreRA_Scheduler scheduler;
-        // scheduler.runOnMachineFunction(mfunc.get());
-
-        // DEBUG = 0;
-        // DEEPDEBUG = 0;
-        // DEBUG = 1;
-        // DEEPDEBUG = 1;
-        // 阶段 3: 物理寄存器分配 (Register Allocation)
-        RISCv64RegAlloc reg_alloc(mfunc.get());
-        reg_alloc.run();
-
-        // DEBUG = 0;
-        // DEEPDEBUG = 0;
-        DEBUG = 1;
-        DEEPDEBUG = 1;
-        if (DEBUG) {
-            std::cerr << "====== stack info after reg alloc ======\n";
-            mfunc->dumpStackFrameInfo(std::cerr);
-        }
-
-        // 阶段 3.1: 处理被调用者保存寄存器
-        CalleeSavedHandler callee_handler;
-        callee_handler.runOnMachineFunction(mfunc.get());
-
-        if (DEBUG) {
-            std::cerr << "====== stack info after callee handler ======\n";
-            mfunc->dumpStackFrameInfo(std::cerr);
-        }
-
-        // // 阶段 4: 窥孔优化 (Peephole Optimization)
-        // PeepholeOptimizer peephole;
-        // peephole.runOnMachineFunction(mfunc.get());
-
-        // 阶段 5: 局部指令调度 (Local Scheduling)
-        // PostRA_Scheduler local_scheduler;
-        // local_scheduler.runOnMachineFunction(mfunc.get());
-
-        // 阶段 3.2: 插入序言和尾声
-        PrologueEpilogueInsertionPass pei_pass;
-        pei_pass.runOnMachineFunction(mfunc.get());
-
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-
-        // 阶段 3.3: 大立即数合法化
-        LegalizeImmediatesPass legalizer;
-        legalizer.runOnMachineFunction(mfunc.get());
-
-        // 阶段 6: 代码发射 (Code Emission)
-        std::stringstream ss;
-        RISCv64AsmPrinter printer(mfunc.get());
-        printer.run(ss);
-
-        if (DEBUG) {
-            ss << "\n\n; --- Intermediate Representation after Instruction Selection ---\n" 
-            << ss_after_isel.str();
-        }
-        DEBUG = 1;
-        DEEPDEBUG = 1;
-        return ss.str();
+        std::cout << ss_after_isel.str();
+    }
+    if (DEBUG) {
+        std::cerr << "====== Intermediate Representation after Instruction Selection ======\n" 
+        << ss_after_isel.str();
     }
     
-    
-    
-    
-    
-    
-    
-    
-    
-    
-    
-    
-    else {
-        // === 完整的后端处理流水线 ===
+    // 阶段 2: 消除帧索引 (展开伪指令，计算局部变量偏移)
+    // 这个Pass必须在寄存器分配之前运行
+    EliminateFrameIndicesPass efi_pass;
+    efi_pass.runOnMachineFunction(mfunc.get());
 
-        // 阶段 1: 指令选择 (sysy::IR -> LLIR with virtual registers)
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-
-        RISCv64ISel isel;
-        std::unique_ptr<MachineFunction> mfunc = isel.runOnFunction(func);
-
-        // 第一次调试打印输出
-        std::stringstream ss_after_isel;
-        RISCv64AsmPrinter printer_isel(mfunc.get());
-        printer_isel.run(ss_after_isel, true);
-        // if (DEBUG) {
-        //     std::cout << ss_after_isel.str();
-        // }
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-        // DEBUG = 1;
-        // DEEPDEBUG = 1;
-        if (DEBUG) {
-            std::cerr << "====== Intermediate Representation after Instruction Selection ======\n" 
-            << ss_after_isel.str();
-        }
-        // DEBUG = 0;
-        // DEEPDEBUG = 0;
-        // [新增] 阶段 2: 消除帧索引 (展开伪指令，计算局部变量偏移)
-        // 这个Pass必须在寄存器分配之前运行
-        EliminateFrameIndicesPass efi_pass;
-        efi_pass.runOnMachineFunction(mfunc.get());
-
-        if (DEBUG) {
-            std::cerr << "====== stack info after eliminate frame indices  ======\n";
-            mfunc->dumpStackFrameInfo(std::cerr);
-            std::stringstream ss_after_eli;
-            printer_isel.run(ss_after_eli, true);
-            std::cerr << "====== LLIR after eliminate frame indices ======\n" 
-            << ss_after_eli.str();
-        }
-
-        // // 阶段 2: 指令调度 (Instruction Scheduling)
-        // PreRA_Scheduler scheduler;
-        // scheduler.runOnMachineFunction(mfunc.get());
-
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-        // DEBUG = 1;
-        // DEEPDEBUG = 1;
-        // 阶段 3: 物理寄存器分配 (Register Allocation)
-        RISCv64RegAlloc reg_alloc(mfunc.get());
-        reg_alloc.run();
-
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-        // DEBUG = 1;
-        // DEEPDEBUG = 1;
-        if (DEBUG) {
-            std::cerr << "====== stack info after reg alloc ======\n";
-            mfunc->dumpStackFrameInfo(std::cerr);
-        }
-
-        // 阶段 3.1: 处理被调用者保存寄存器
-        CalleeSavedHandler callee_handler;
-        callee_handler.runOnMachineFunction(mfunc.get());
-
-        if (DEBUG) {
-            std::cerr << "====== stack info after callee handler ======\n";
-            mfunc->dumpStackFrameInfo(std::cerr);
-        }
-
-        // // 阶段 4: 窥孔优化 (Peephole Optimization)
-        // PeepholeOptimizer peephole;
-        // peephole.runOnMachineFunction(mfunc.get());
-
-        // 阶段 5: 局部指令调度 (Local Scheduling)
-        // PostRA_Scheduler local_scheduler;
-        // local_scheduler.runOnMachineFunction(mfunc.get());
-
-        // 阶段 3.2: 插入序言和尾声
-        PrologueEpilogueInsertionPass pei_pass;
-        pei_pass.runOnMachineFunction(mfunc.get());
-
-        DEBUG = 0;
-        DEEPDEBUG = 0;
-
-        // 阶段 3.3: 大立即数合法化
-        LegalizeImmediatesPass legalizer;
-        legalizer.runOnMachineFunction(mfunc.get());
-
-        // 阶段 6: 代码发射 (Code Emission)
-        std::stringstream ss;
-        RISCv64AsmPrinter printer(mfunc.get());
-        printer.run(ss);
-
-        if (DEBUG) {
-            ss << "\n\n; --- Intermediate Representation after Instruction Selection ---\n" 
-            << ss_after_isel.str();
-        }
-        return ss.str();
+    if (DEBUG) {
+        std::cerr << "====== stack info after eliminate frame indices  ======\n";
+        mfunc->dumpStackFrameInfo(std::cerr);
+        std::stringstream ss_after_eli;
+        printer_isel.run(ss_after_eli, true);
+        std::cerr << "====== LLIR after eliminate frame indices ======\n" 
+        << ss_after_eli.str();
     }
+
+    // 阶段 2: 除法强度削弱优化 (Division Strength Reduction)
+    DivStrengthReduction div_strength_reduction;
+    div_strength_reduction.runOnMachineFunction(mfunc.get());
+
+    // 阶段 2.1: 指令调度 (Instruction Scheduling)
+    PreRA_Scheduler scheduler;
+    scheduler.runOnMachineFunction(mfunc.get());
+
+    // 阶段 3: 物理寄存器分配 (Register Allocation)
+    RISCv64RegAlloc reg_alloc(mfunc.get());
+    reg_alloc.run();
+
+    if (DEBUG) {
+        std::cerr << "====== stack info after reg alloc ======\n";
+        mfunc->dumpStackFrameInfo(std::cerr);
+    }
+
+    // 阶段 3.1: 处理被调用者保存寄存器
+    CalleeSavedHandler callee_handler;
+    callee_handler.runOnMachineFunction(mfunc.get());
+
+    if (DEBUG) {
+        std::cerr << "====== stack info after callee handler ======\n";
+        mfunc->dumpStackFrameInfo(std::cerr);
+    }
+
+    // 阶段 4: 窥孔优化 (Peephole Optimization)
+    PeepholeOptimizer peephole;
+    peephole.runOnMachineFunction(mfunc.get());
+
+    // 阶段 5: 局部指令调度 (Local Scheduling)
+    PostRA_Scheduler local_scheduler;
+    local_scheduler.runOnMachineFunction(mfunc.get());
+
+    // 阶段 3.2: 插入序言和尾声
+    PrologueEpilogueInsertionPass pei_pass;
+    pei_pass.runOnMachineFunction(mfunc.get());
+
+    // 阶段 3.3: 大立即数合法化
+    LegalizeImmediatesPass legalizer;
+    legalizer.runOnMachineFunction(mfunc.get());
+
+    // 阶段 6: 代码发射 (Code Emission)
+    std::stringstream ss;
+    RISCv64AsmPrinter printer(mfunc.get());
+    printer.run(ss);
+
+    return ss.str();
+    
 }
 
 } // namespace sysy

src/backend/RISCv64/RISCv64ISel.cpp

@@ -1,9 +1,10 @@
 #include "RISCv64ISel.h"
+#include "IR.h"  // For GlobalValue
 #include <stdexcept>
 #include <set>
 #include <functional>
-#include <cmath> // For std::fabs
-#include <limits> // For std::numeric_limits
+#include <cmath>
+#include <limits>
 #include <iostream>
 
 namespace sysy {
@@ -167,33 +168,6 @@ void RISCv64ISel::selectBasicBlock(BasicBlock* bb) {
             select_recursive(node_to_select);
         }
     }
-
-    if (CurMBB == MFunc->getBlocks().front().get()) { // 只对入口块操作
-        auto keepalive = std::make_unique<MachineInstr>(RVOpcodes::PSEUDO_KEEPALIVE);
-        for (Argument* arg : F->getArguments()) {
-            keepalive->addOperand(std::make_unique<RegOperand>(getVReg(arg)));
-        }
-
-        auto& instrs = CurMBB->getInstructions();
-        auto insert_pos = instrs.end();
-
-        // 关键：检查基本块是否以一个“终止指令”结尾
-        if (!instrs.empty()) {
-            RVOpcodes last_op = instrs.back()->getOpcode();
-            // 扩充了判断条件，涵盖所有可能的终止指令
-            if (last_op == RVOpcodes::J || last_op == RVOpcodes::RET ||
-                last_op == RVOpcodes::BEQ || last_op == RVOpcodes::BNE ||
-                last_op == RVOpcodes::BLT || last_op == RVOpcodes::BGE ||
-                last_op == RVOpcodes::BLTU || last_op == RVOpcodes::BGEU)
-            {
-                // 如果是，插入点就在这个终止指令之前
-                insert_pos = std::prev(instrs.end());
-            }
-        }
-        
-        // 在计算出的正确位置插入伪指令
-        instrs.insert(insert_pos, std::move(keepalive));
-    }
 }
 
 // 核心函数：为DAG节点选择并生成MachineInstr (已修复和增强的完整版本)
@@ -209,8 +183,12 @@ void RISCv64ISel::selectNode(DAGNode* node) {
         case DAGNode::CONSTANT:
         case DAGNode::ALLOCA_ADDR:
             if (node->value) {
-                // 确保它有一个关联的虚拟寄存器即可，不生成代码。
-                getVReg(node->value);
+                // GlobalValue objects (global variables) should not get virtual registers
+                // since they represent memory addresses, not register-allocated values
+                if (dynamic_cast<GlobalValue*>(node->value) == nullptr) {
+                    // 确保它有一个关联的虚拟寄存器即可，不生成代码。
+                    getVReg(node->value);
+                }
             }
             break;
         
@@ -402,7 +380,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                 Value* base = nullptr;
                 Value* offset = nullptr;
 
-                // [修改] 扩展基地址的判断，使其可以识别 AllocaInst 或 GlobalValue
+                // 扩展基地址的判断，使其可以识别 AllocaInst 或 GlobalValue
                 if (dynamic_cast<AllocaInst*>(lhs) || dynamic_cast<GlobalValue*>(lhs)) {
                     base = lhs;
                     offset = rhs;
@@ -421,7 +399,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                         CurMBB->addInstruction(std::move(li));
                     }
                     
-                    // 2. [修改] 根据基地址的类型，生成不同的指令来获取基地址
+                    // 2. 根据基地址的类型，生成不同的指令来获取基地址
                     auto base_addr_vreg = getNewVReg(Type::getIntType()); // 创建一个新的临时vreg来存放基地址
 
                     // 情况一：基地址是局部栈变量
@@ -452,7 +430,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                 }
             }
 
-            // [V2优点] 在BINARY节点内部按需加载常量操作数。
+            // 在BINARY节点内部按需加载常量操作数。
             auto load_val_if_const = [&](Value* val) {
                 if (auto c = dynamic_cast<ConstantValue*>(val)) {
                     if (DEBUG) {
@@ -483,7 +461,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
             auto dest_vreg = getVReg(bin);
             auto lhs_vreg = getVReg(lhs);
 
-            // [V2优点] 融合 ADDIW 优化。
+            // 融合 ADDIW 优化。
             if (rhs_is_imm_opt) {
                 auto rhs_const = dynamic_cast<ConstantValue*>(rhs);
                 auto instr = std::make_unique<MachineInstr>(RVOpcodes::ADDIW);
@@ -539,6 +517,15 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                     CurMBB->addInstruction(std::move(instr));
                     break;
                 }
+                case Instruction::kSra: {
+                    auto rhs_const = dynamic_cast<ConstantInteger*>(rhs);
+                    auto instr = std::make_unique<MachineInstr>(RVOpcodes::SRAIW);
+                    instr->addOperand(std::make_unique<RegOperand>(dest_vreg));
+                    instr->addOperand(std::make_unique<RegOperand>(lhs_vreg));
+                    instr->addOperand(std::make_unique<ImmOperand>(rhs_const->getInt()));
+                    CurMBB->addInstruction(std::move(instr));
+                    break;
+                }
                 case BinaryInst::kICmpEQ: { // 等于 (a == b) -> (subw; seqz)
                     auto sub = std::make_unique<MachineInstr>(RVOpcodes::SUBW);
                     sub->addOperand(std::make_unique<RegOperand>(dest_vreg));
@@ -758,11 +745,83 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                     CurMBB->addInstruction(std::move(instr));
                     break;
                 }
-                case Instruction::kFtoI: { // 浮点 to 整数
-                    auto instr = std::make_unique<MachineInstr>(RVOpcodes::FCVT_W_S);
-                    instr->addOperand(std::make_unique<RegOperand>(dest_vreg)); // 目标是整数vreg
-                    instr->addOperand(std::make_unique<RegOperand>(src_vreg));  // 源是浮点vreg
-                    CurMBB->addInstruction(std::move(instr));
+                case Instruction::kFtoI: { // 浮点 to 整数 (带向下取整)
+                    // 目标：实现 floor(x) 的效果, C/C++中浮点转整数是截断(truncate)
+                    // 对于正数，floor(x) == truncate(x)
+                    // RISC-V的 fcvt.w.s 默认是“四舍五入到偶数”
+                    // 我们需要手动实现截断逻辑
+                    // 逻辑:
+                    //   temp_i = fcvt.w.s(x)  // 四舍五入
+                    //   temp_f = fcvt.s.w(temp_i) // 转回浮点
+                    //   if (x < temp_f) { // 如果原数更小，说明被“五入”了
+                    //     result = temp_i - 1
+                    //   } else {
+                    //     result = temp_i
+                    //   }
+                    
+                    auto temp_i_vreg = getNewVReg(Type::getIntType());
+                    auto temp_f_vreg = getNewVReg(Type::getFloatType());
+                    auto cmp_vreg = getNewVReg(Type::getIntType());
+
+                    // 1. fcvt.w.s temp_i_vreg, src_vreg
+                    auto fcvt_w = std::make_unique<MachineInstr>(RVOpcodes::FCVT_W_S);
+                    fcvt_w->addOperand(std::make_unique<RegOperand>(temp_i_vreg));
+                    fcvt_w->addOperand(std::make_unique<RegOperand>(src_vreg));
+                    CurMBB->addInstruction(std::move(fcvt_w));
+
+                    // 2. fcvt.s.w temp_f_vreg, temp_i_vreg
+                    auto fcvt_s = std::make_unique<MachineInstr>(RVOpcodes::FCVT_S_W);
+                    fcvt_s->addOperand(std::make_unique<RegOperand>(temp_f_vreg));
+                    fcvt_s->addOperand(std::make_unique<RegOperand>(temp_i_vreg));
+                    CurMBB->addInstruction(std::move(fcvt_s));
+
+                    // 3. flt.s cmp_vreg, src_vreg, temp_f_vreg
+                    auto flt = std::make_unique<MachineInstr>(RVOpcodes::FLT_S);
+                    flt->addOperand(std::make_unique<RegOperand>(cmp_vreg));
+                    flt->addOperand(std::make_unique<RegOperand>(src_vreg));
+                    flt->addOperand(std::make_unique<RegOperand>(temp_f_vreg));
+                    CurMBB->addInstruction(std::move(flt));
+
+                    // 创建标签
+                    int unique_id = this->local_label_counter++;
+                    std::string rounded_up_label = MFunc->getName() + "_ftoi_rounded_up_" + std::to_string(unique_id);
+                    std::string done_label = MFunc->getName() + "_ftoi_done_" + std::to_string(unique_id);
+
+                    // 4. bne cmp_vreg, x0, rounded_up_label
+                    auto bne = std::make_unique<MachineInstr>(RVOpcodes::BNE);
+                    bne->addOperand(std::make_unique<RegOperand>(cmp_vreg));
+                    bne->addOperand(std::make_unique<RegOperand>(PhysicalReg::ZERO));
+                    bne->addOperand(std::make_unique<LabelOperand>(rounded_up_label));
+                    CurMBB->addInstruction(std::move(bne));
+
+                    // 5. else 分支: mv dest_vreg, temp_i_vreg
+                    auto mv = std::make_unique<MachineInstr>(RVOpcodes::MV);
+                    mv->addOperand(std::make_unique<RegOperand>(dest_vreg));
+                    mv->addOperand(std::make_unique<RegOperand>(temp_i_vreg));
+                    CurMBB->addInstruction(std::move(mv));
+
+                    // 6. j done_label
+                    auto j = std::make_unique<MachineInstr>(RVOpcodes::J);
+                    j->addOperand(std::make_unique<LabelOperand>(done_label));
+                    CurMBB->addInstruction(std::move(j));
+
+                    // 7. rounded_up_label:
+                    auto label_up = std::make_unique<MachineInstr>(RVOpcodes::LABEL);
+                    label_up->addOperand(std::make_unique<LabelOperand>(rounded_up_label));
+                    CurMBB->addInstruction(std::move(label_up));
+
+                    // 8. addiw dest_vreg, temp_i_vreg, -1
+                    auto addi = std::make_unique<MachineInstr>(RVOpcodes::ADDIW);
+                    addi->addOperand(std::make_unique<RegOperand>(dest_vreg));
+                    addi->addOperand(std::make_unique<RegOperand>(temp_i_vreg));
+                    addi->addOperand(std::make_unique<ImmOperand>(-1));
+                    CurMBB->addInstruction(std::move(addi));
+
+                    // 9. done_label:
+                    auto label_done = std::make_unique<MachineInstr>(RVOpcodes::LABEL);
+                    label_done->addOperand(std::make_unique<LabelOperand>(done_label));
+                    CurMBB->addInstruction(std::move(label_done));
+                    
                     break;
                 }
                 case Instruction::kFNeg: { // 浮点取负
@@ -943,7 +1002,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
             
             // --- 步骤 3: 生成CALL指令 ---
             auto call_instr = std::make_unique<MachineInstr>(RVOpcodes::CALL);
-            // [协议] 如果函数有返回值，将它的目标虚拟寄存器作为第一个操作数
+            // 如果函数有返回值，将它的目标虚拟寄存器作为第一个操作数
             if (!call->getType()->isVoid()) {
                 unsigned dest_vreg = getVReg(call);
                 call_instr->addOperand(std::make_unique<RegOperand>(dest_vreg));
@@ -1020,7 +1079,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                 } else {
                     // --- 处理整数/指针返回值 ---
                     // 返回值需要被放入 a0
-                    // [V2优点] 在RETURN节点内加载常量返回值
+                    // 在RETURN节点内加载常量返回值
                     if (auto const_val = dynamic_cast<ConstantValue*>(ret_val)) {
                         auto li_instr = std::make_unique<MachineInstr>(RVOpcodes::LI);
                         li_instr->addOperand(std::make_unique<RegOperand>(PhysicalReg::A0));
@@ -1034,7 +1093,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                     }
                 }
             }
-            // [V1设计保留] 函数尾声（epilogue）不由RETURN节点生成，
+            // 函数尾声（epilogue）不由RETURN节点生成，
             // 而是由后续的AsmPrinter或其它Pass统一处理，这是一种常见且有效的模块化设计。
             auto ret_mi = std::make_unique<MachineInstr>(RVOpcodes::RET);
             CurMBB->addInstruction(std::move(ret_mi));
@@ -1048,7 +1107,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
             auto then_bb_name = cond_br->getThenBlock()->getName();
             auto else_bb_name = cond_br->getElseBlock()->getName();
 
-            // [优化] 检查分支条件是否为编译期常量
+            // 检查分支条件是否为编译期常量
             if (auto const_cond = dynamic_cast<ConstantValue*>(condition)) {
                 // 如果条件是常量，直接生成一个无条件跳转J，而不是BNE
                 if (const_cond->getInt() != 0) { // 条件为 true
@@ -1063,7 +1122,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
             } 
             // 如果条件不是常量，则执行标准流程
             else {
-                // [修复] 为条件变量生成加载指令（如果它是常量的话，尽管上面已经处理了）
+                // 为条件变量生成加载指令（如果它是常量的话，尽管上面已经处理了）
                 // 这一步是为了逻辑完整，以防有其他类型的常量没有被捕获
                 if (auto const_val = dynamic_cast<ConstantValue*>(condition)) {
                     auto li = std::make_unique<MachineInstr>(RVOpcodes::LI);
@@ -1097,7 +1156,7 @@ void RISCv64ISel::selectNode(DAGNode* node) {
     }
 
         case DAGNode::MEMSET: {
-            // [V1设计保留] Memset的核心展开逻辑在虚拟寄存器层面是正确的，无需修改。
+            // Memset的核心展开逻辑在虚拟寄存器层面是正确的，无需修改。
             // 之前的bug是由于其输入（地址、值、大小）的虚拟寄存器未被正确初始化。
             // 在修复了CONSTANT/ALLOCA_ADDR的加载问题后，此处的逻辑现在可以正常工作。
 
@@ -1280,14 +1339,19 @@ void RISCv64ISel::selectNode(DAGNode* node) {
                 if (stride != 0) {
                     // --- 为当前索引和步长生成偏移计算指令 ---
                     auto offset_vreg = getNewVReg();
-                    auto index_vreg = getVReg(indexValue);
-
-                    // 如果索引是常量，先用 LI 指令加载到虚拟寄存器
+                    
+                    // 处理索引 - 区分常量与动态值
+                    unsigned index_vreg;
                     if (auto const_index = dynamic_cast<ConstantValue*>(indexValue)) {
+                        // 对于常量索引，直接创建新的虚拟寄存器
+                        index_vreg = getNewVReg();
                         auto li = std::make_unique<MachineInstr>(RVOpcodes::LI);
                         li->addOperand(std::make_unique<RegOperand>(index_vreg));
                         li->addOperand(std::make_unique<ImmOperand>(const_index->getInt()));
                         CurMBB->addInstruction(std::move(li));
+                    } else {
+                        // 对于动态索引，使用已存在的虚拟寄存器
+                        index_vreg = getVReg(indexValue);
                     }
                     
                     // 优化：如果步长是1，可以直接移动(MV)作为偏移量，无需乘法
@@ -1445,7 +1509,7 @@ std::vector<std::unique_ptr<RISCv64ISel::DAGNode>> RISCv64ISel::build_dag(BasicB
             
             // 依次添加所有索引作为后续的操作数
             for (auto index : gep->getIndices()) {
-                // [修复] 从 Use 对象中获取真正的 Value*
+                // 从 Use 对象中获取真正的 Value*
                 gep_node->operands.push_back(get_operand_node(index->getValue(), value_to_node, nodes_storage));
             }
         } else if (auto load = dynamic_cast<LoadInst*>(inst)) {
@@ -1473,7 +1537,7 @@ std::vector<std::unique_ptr<RISCv64ISel::DAGNode>> RISCv64ISel::build_dag(BasicB
                     }
                 }
             }
-            if (bin->getKind() >= Instruction::kFAdd) { // 假设浮点指令枚举值更大
+            if (bin->isFPBinary()) { // 假设浮点指令枚举值更大
                 auto fbin_node = create_node(DAGNode::FBINARY, bin, value_to_node, nodes_storage);
                 fbin_node->operands.push_back(get_operand_node(bin->getLhs(), value_to_node, nodes_storage));
                 fbin_node->operands.push_back(get_operand_node(bin->getRhs(), value_to_node, nodes_storage));
@@ -1549,7 +1613,7 @@ unsigned RISCv64ISel::getTypeSizeInBytes(Type* type) {
     }
 }
 
-// [新] 打印DAG图以供调试的辅助函数
+// 打印DAG图以供调试的辅助函数
 void RISCv64ISel::print_dag(const std::vector<std::unique_ptr<DAGNode>>& dag, const std::string& bb_name) {
     // 检查是否有DEBUG宏或者全局变量，避免在非调试模式下打印
     // if (!DEBUG) return; 
@@ -1645,4 +1709,8 @@ void RISCv64ISel::print_dag(const std::vector<std::unique_ptr<DAGNode>>& dag, co
     std::cerr << "======================================\n\n";
 }
 
+unsigned int RISCv64ISel::getVRegCounter() const {
+    return vreg_counter;
+}
+
 } // namespace sysy

src/backend/RISCv64/RISCv64LLIR.cpp

@@ -44,7 +44,7 @@ std::string regToString(PhysicalReg reg) {
     }
 }
 
-// [新增] 打印栈帧信息的完整实现
+// 打印栈帧信息的完整实现
 void MachineFunction::dumpStackFrameInfo(std::ostream& os) const {
     const StackFrameInfo& info = frame_info;
 

src/backend/RISCv64/RISCv64RegAlloc.cpp

@@ -55,12 +55,41 @@ void RISCv64RegAlloc::run() {
 
     if (DEBUG) std::cerr << "===== Running Graph Coloring Register Allocation for function: " << MFunc->getName() << " =====\n";
     
-    while (true) {
+    const int MAX_ITERATIONS = 50;
+    int iteration = 0;
+    
+    while (iteration++ < MAX_ITERATIONS) {
         if (doAllocation()) {
             break;
         } else {
             rewriteProgram();
-            if (DEBUG) std::cerr << "--- Spilling detected, re-running allocation ---\n";
+            if (DEBUG) std::cerr << "--- Spilling detected, re-running allocation (iteration " << iteration << ") ---\n";
+            
+            if (iteration >= MAX_ITERATIONS) {
+                std::cerr << "ERROR: Register allocation failed to converge after " << MAX_ITERATIONS << " iterations\n";
+                std::cerr << "  Spill worklist size: " << spillWorklist.size() << "\n";
+                std::cerr << "  Total nodes: " << (initial.size() + coloredNodes.size()) << "\n";
+                
+                // Emergency spill remaining nodes to break the loop
+                std::cerr << "  Emergency spilling remaining spill worklist nodes...\n";
+                for (unsigned node : spillWorklist) {
+                    spilledNodes.insert(node);
+                }
+                
+                // Also spill any nodes that didn't get colors
+                std::set<unsigned> uncolored;
+                for (unsigned node : initial) {
+                    if (color_map.find(node) == color_map.end()) {
+                        uncolored.insert(node);
+                    }
+                }
+                for (unsigned node : uncolored) {
+                    spilledNodes.insert(node);
+                }
+                
+                // Force completion
+                break;
+            }
         }
     }
 
@@ -93,7 +122,7 @@ bool RISCv64RegAlloc::doAllocation() {
 }
 
 void RISCv64RegAlloc::precolorByCallingConvention() {
-    // [新增] 在处理前，先清空颜色相关的状态，确保重试时不会出错
+    // 在处理前，先清空颜色相关的状态，确保重试时不会出错
     color_map.clear();
     coloredNodes.clear();
 
@@ -122,30 +151,6 @@ void RISCv64RegAlloc::precolorByCallingConvention() {
         }
     }
 
-    // // --- 部分2：为CALL指令的返回值预着色 ---
-    // for (auto& mbb : MFunc->getBlocks()) {
-    //     for (auto& instr : mbb->getInstructions()) {
-    //         if (instr->getOpcode() == RVOpcodes::CALL) {
-    //             if (!instr->getOperands().empty() && 
-    //                 instr->getOperands().front()->getKind() == MachineOperand::KIND_REG) 
-    //             {
-    //                 auto reg_op = static_cast<RegOperand*>(instr->getOperands().front().get());
-    //                 if (reg_op->isVirtual()) {
-    //                     unsigned ret_vreg = reg_op->getVRegNum();
-    //                     assert(vreg_to_value_map.count(ret_vreg) && "Return vreg not found!");
-    //                     Value* ret_val = vreg_to_value_map.at(ret_vreg);
-
-    //                     if (ret_val->getType()->isFloat()) {
-    //                         color_map[ret_vreg] = PhysicalReg::F10; // fa0
-    //                     } else {
-    //                         color_map[ret_vreg] = PhysicalReg::A0;  // a0
-    //                     }
-    //                 }
-    //             }
-    //         }
-    //     }
-    // }
-    
     // 将所有预着色的vreg视为已着色节点
     for(const auto& pair : color_map) {
         coloredNodes.insert(pair.first);
@@ -168,7 +173,7 @@ void RISCv64RegAlloc::initialize() {
     spillWorklist.clear();
     spilledNodes.clear();
     coalescedNodes.clear();
-    coloredNodes.clear(); // [修正] 恢复对这两个集合的清除
+    coloredNodes.clear();
     selectStack.clear();
 
     coalescedMoves.clear();
@@ -181,7 +186,7 @@ void RISCv64RegAlloc::initialize() {
     degree.clear();
     moveList.clear();
     alias.clear();
-    color_map.clear(); // [修正] 恢复对 color_map 的清除
+    color_map.clear();
 }
 
 // 活跃性分析（此部分为标准数据流分析，与现有版本类似但更精细）
@@ -197,7 +202,6 @@ void RISCv64RegAlloc::analyzeLiveness() {
         VRegSet uses, defs;
         for (const auto& instr_ptr : mbb->getInstructions()) {
             VRegSet instr_use, instr_def;
-            // 使用新的、能看到物理寄存器的版本
             getInstrUseDef_Liveness(instr_ptr.get(), instr_use, instr_def);
             for (unsigned u : instr_use) {
                 if (defs.find(u) == defs.end()) uses.insert(u);
@@ -238,7 +242,7 @@ void RISCv64RegAlloc::analyzeLiveness() {
             const MachineInstr* instr = instr_it->get();
             live_out_map[instr] = current_live;
             VRegSet use, def;
-            // 使用新的、能看到物理寄存器的版本
+
             getInstrUseDef_Liveness(instr, use, def);
             for(auto d : def) current_live.erase(d);
             for(auto u : use) current_live.insert(u);
@@ -247,7 +251,6 @@ void RISCv64RegAlloc::analyzeLiveness() {
     }
 }
 
-// [最终修正] 包含了所有正确逻辑和您已有调试代码的完整版本
 void RISCv64RegAlloc::build() {
     initial.clear(); 
     RISCv64AsmPrinter printer_inside_build(MFunc);
@@ -403,15 +406,33 @@ void RISCv64RegAlloc::build() {
             }
             
             // --- 规则 3: Live_Out 集合内部的【虚拟寄存器】形成完全图 ---
-            // [优化与修正] 使用更高效的遍历，避免重复调用 addEdge(A,B) 和 addEdge(B,A)
-            for (auto it1 = live_out.begin(); it1 != live_out.end(); ++it1) {
-                unsigned l1 = *it1;
-                // [关键修正] 只为虚拟寄存器 l1 添加边
-                if (precolored.count(l1)) continue;
+            // 使用更高效的遍历，避免重复调用 addEdge(A,B) 和 addEdge(B,A)
+            // 添加限制以防止过度密集的图
+            const size_t MAX_LIVE_OUT_SIZE = 32; // 限制最大活跃变量数
+            if (live_out.size() > MAX_LIVE_OUT_SIZE) {
+                // 对于大量活跃变量，使用更保守的边添加策略
+                // 只添加必要的边，而不是完全图
+                for (auto it1 = live_out.begin(); it1 != live_out.end(); ++it1) {
+                    unsigned l1 = *it1;
+                    if (precolored.count(l1)) continue;
+                    
+                    // 只添加与定义变量相关的边
+                    for (unsigned d : def) {
+                        if (d != l1 && !precolored.count(d)) {
+                            addEdge(l1, d);
+                        }
+                    }
+                }
+            } else {
+                // 对于较小的集合，使用原来的完全图方法
+                for (auto it1 = live_out.begin(); it1 != live_out.end(); ++it1) {
+                    unsigned l1 = *it1;
+                    if (precolored.count(l1)) continue;
 
-                for (auto it2 = std::next(it1); it2 != live_out.end(); ++it2) {
-                    unsigned l2 = *it2;
-                    addEdge(l1, l2);
+                    for (auto it2 = std::next(it1); it2 != live_out.end(); ++it2) {
+                        unsigned l2 = *it2;
+                        addEdge(l1, l2);
+                    }
                 }
             }
         }
@@ -463,6 +484,17 @@ void RISCv64RegAlloc::coalesce() {
     unsigned y = getAlias(*use.begin());
     unsigned u, v;
     if (precolored.count(y)) { u = y; v = x; } else { u = x; v = y; }
+    
+    // 防御性检查，处理物理寄存器之间的传送指令
+    if (precolored.count(u) && precolored.count(v)) {
+        // 如果 u 和 v 都是物理寄存器，我们不能合并它们。
+        // 这通常是一条寄存器拷贝指令，例如 `mv a2, a1`。
+        // 把它加入 constrainedMoves 列表，然后直接返回，不再处理。
+        constrainedMoves.insert(move);
+        // addWorklist(u) 和 addWorklist(v) 在这里也不需要调用，
+        // 因为它们只对虚拟寄存器有意义。
+        return;
+    }
 
     if (DEEPERDEBUG) std::cerr << "[Coalesce] Processing move between " << regIdToString(x) 
                            << " and " << regIdToString(y) << " (aliases " << regIdToString(u) 
@@ -484,8 +516,6 @@ void RISCv64RegAlloc::coalesce() {
         addWorklist(v);
         return; // 立即返回，不再进行后续检查
     }
-
-    // --- 新的、拆分后的启发式检查逻辑 ---
     
     bool pre_interfere = adjList.at(v).count(u);
     
@@ -504,8 +534,6 @@ void RISCv64RegAlloc::coalesce() {
         // --- 场景1：u是物理寄存器，使用 George 启发式 ---
         if (DEEPERDEBUG) std::cerr << "  -> Trying George Heuristic (u is precolored)...\n";
         
-        // ==================== [展开的 std::all_of 逻辑] ====================
-        
         // 步骤 1: 独立调用 adjacent(v) 获取邻居集合
         VRegSet neighbors_of_v = adjacent(v);
         if (DEEPERDEBUG) {
@@ -537,7 +565,6 @@ void RISCv64RegAlloc::coalesce() {
         if (DEEPERDEBUG) {
              std::cerr << "  -> George Heuristic final result: " << (george_ok ? "OK" : "FAIL") << "\n";
         }
-        // =================================================================
 
         if (george_ok) {
             can_coalesce = true;
@@ -622,11 +649,11 @@ void RISCv64RegAlloc::assignColors() {
         }
     }
 
-    // 步骤 2: [最终修正] 完整、正确地处理 coalescedNodes
+    // 步骤 2: 处理 coalescedNodes
     for (unsigned n : coalescedNodes) {
         unsigned root_alias = getAlias(n);
         
-        // --- 新的、健壮的逻辑，处理所有三种可能性 ---
+        // --- 处理所有三种可能性 ---
 
         // 情况 1: 别名本身就是物理寄存器 (修复当前bug)
         if (precolored.count(root_alias)) {
@@ -649,9 +676,13 @@ void RISCv64RegAlloc::assignColors() {
 
 // 重写程序，插入溢出代码
 void RISCv64RegAlloc::rewriteProgram() {
-    // 1. 为溢出的旧vreg在栈上分配空间
     StackFrameInfo& frame_info = MFunc->getFrameInfo();
-    int spill_base_offset = frame_info.locals_size + frame_info.callee_saved_size;
+    // 使用 EFI Pass 确定的 locals_end_offset 作为溢出分配的基准。
+    // locals_end_offset 本身是负数，代表局部变量区域的下边界地址。
+    int spill_current_offset = frame_info.locals_end_offset - frame_info.spill_size;
+
+    // 保存溢出区域的起始点，用于最后计算总的 spill_size
+    const int spill_start_offset = frame_info.locals_end_offset;
 
     for (unsigned vreg : spilledNodes) {
         if (frame_info.spill_offsets.count(vreg)) continue;
@@ -663,11 +694,19 @@ void RISCv64RegAlloc::rewriteProgram() {
             size = 8; // pointer
         }
 
-        spill_base_offset += size;
-        spill_base_offset = (spill_base_offset + 7) & ~7;
-        frame_info.spill_offsets[vreg] = -spill_base_offset; // [修正] 溢出槽也使用负偏移量
+        // 在当前偏移基础上继续向下(地址变得更负)分配空间
+        spill_current_offset -= size;
+        
+        // 对齐新的、更小的地址，RISC-V 要求8字节对齐
+        spill_current_offset = spill_current_offset & ~7;
+
+        // 将计算出的、不会冲突的正确偏移量存入 spill_offsets
+        frame_info.spill_offsets[vreg] = spill_current_offset;
     }
-    frame_info.spill_size = spill_base_offset - (frame_info.locals_size + frame_info.callee_saved_size);
+
+    // 更新总的溢出区域大小。
+    // spill_size = -(结束偏移 - 开始偏移)
+    frame_info.spill_size = -(spill_current_offset - spill_start_offset);
 
     // 2. 遍历所有指令，重写代码
     for (auto& mbb : MFunc->getBlocks()) {
@@ -711,7 +750,7 @@ void RISCv64RegAlloc::rewriteProgram() {
                 }
             }
 
-            // c. [核心修正] 创建一条全新的指令，用新vreg替换旧vreg
+            // c. 创建一条全新的指令，用新vreg替换旧vreg
             auto new_instr = std::make_unique<MachineInstr>(instr_ptr->getOpcode());
             for (const auto& op : instr_ptr->getOperands()) {
                 if (op->getKind() == MachineOperand::KIND_REG) {
@@ -771,6 +810,11 @@ void RISCv64RegAlloc::rewriteProgram() {
                     std::make_unique<ImmOperand>(frame_info.spill_offsets.at(old_vreg))
                 ));
                 new_instructions.push_back(std::move(store));
+                if (DEEPERDEBUG) {
+                    std::cerr << "[Spill] Inserted spill store for %vreg" << old_vreg 
+                              << " at offset " << frame_info.spill_offsets.at(old_vreg) 
+                              << " with new temp vreg " << regIdToString(new_temp_vreg) << ".\n";
+                }
             }
         }
         mbb->getInstructions() = std::move(new_instructions);
@@ -781,7 +825,7 @@ void RISCv64RegAlloc::rewriteProgram() {
 }
 
 /**
- * @brief [最终修正] 获取一条指令完整的【虚拟】使用/定义寄存器集合
+ * @brief 获取一条指令完整的【虚拟】使用/定义寄存器集合
  * 这个函数将服务于图的构建（收集initial节点等）。
  */
 void RISCv64RegAlloc::getInstrUseDef(const MachineInstr* instr, VRegSet& use, VRegSet& def) {
@@ -832,10 +876,8 @@ void RISCv64RegAlloc::getInstrUseDef(const MachineInstr* instr, VRegSet& use, VR
     }
 }
 
-// in file: RISCv64RegAlloc.cpp
-
 /**
- * @brief [最终修复版] 获取一条指令完整的、包含物理寄存器的Use/Def集合
+ * @brief 获取一条指令完整的、包含物理寄存器的Use/Def集合
  * 这个函数专门服务于活跃性分析，现已补全所有指令（包括伪指令）的逻辑。
  */
 void RISCv64RegAlloc::getInstrUseDef_Liveness(const MachineInstr* instr, VRegSet& use, VRegSet& def) {
@@ -943,7 +985,6 @@ void RISCv64RegAlloc::getInstrUseDef_Liveness(const MachineInstr* instr, VRegSet
     else if (opcode == RVOpcodes::JALR) {
         // JALR rd, rs1, imm. Def: rd, Use: rs1. 
         // 同时也隐式定义了ra(x1)，但通常rd就是ra。为精确，我们只处理显式操作数。
-        // 旧版本逻辑：def.insert(ra); first_reg_is_def = false; -> 这是不精确的
         def.insert(get_any_reg_id(operands[0].get()));
         use.insert(get_any_reg_id(operands[1].get()));
     }
@@ -952,7 +993,7 @@ void RISCv64RegAlloc::getInstrUseDef_Liveness(const MachineInstr* instr, VRegSet
         use.insert(offset + static_cast<unsigned>(PhysicalReg::A0));
         use.insert(offset + static_cast<unsigned>(PhysicalReg::F10)); // F10 is fa0
     }
-    // [关键Bug修复] 添加对 PSEUDO_KEEPALIVE 的处理
+    // 添加对 PSEUDO_KEEPALIVE 的处理
     else if (opcode == RVOpcodes::PSEUDO_KEEPALIVE) {
         // keepalive的所有操作数都是use，以确保它们的生命周期延续到该点
         for (const auto& op : operands) {
@@ -1212,7 +1253,7 @@ bool RISCv64RegAlloc::briggsHeuristic(unsigned u, unsigned v) {
 
 // George启发式
 bool RISCv64RegAlloc::georgeHeuristic(unsigned t, unsigned u) {
-    // 关键修正：如果 t 不是一个待分配的虚拟寄存器（即它是物理寄存器），
+    // 如果 t 不是一个待分配的虚拟寄存器（即它是物理寄存器），
     // 那么它已经被预着色，总是满足 George 启发式条件。
     // 我们通过检查 degree.count(t) 来判断 t 是否在我们的虚拟寄存器工作集中。
     if (degree.count(t) == 0) {
@@ -1224,44 +1265,6 @@ bool RISCv64RegAlloc::georgeHeuristic(unsigned t, unsigned u) {
     return degree.at(t) < K || precolored.count(u) || adjList.at(t).count(u);
 }
 
-// void RISCv64RegAlloc::combine(unsigned u, unsigned v) {
-//     // 1. 从相应的工作列表中移除即将被合并的节点 v
-//     if (freezeWorklist.count(v)) {
-//         freezeWorklist.erase(v);
-//     } else {
-//         spillWorklist.erase(v);
-//     }
-
-//     // 2. 将 v 加入 coalescedNodes 集合，并设置其别名
-//     coalescedNodes.insert(v);
-//     alias[v] = u;
-
-//     // 3. 将 v 的传送指令列表合并到 u
-//     if (moveList.count(u) && moveList.count(v)) {
-//         moveList.at(u).insert(moveList.at(v).begin(), moveList.at(v).end());
-//     } else if (moveList.count(v)) {
-//         moveList[u] = moveList.at(v);
-//     }
-    
-//     // [Bug修复] 移除了论文伪代码中不存在的 enableMoves({v}) 调用。
-
-//     // 4. [核心Bug修复] 遍历 v 的“当前有效”邻居 t (使用 adjacent(v) 而非 adjList.at(v))
-//     //    将它们与 u 连接，并更新它们的度数。
-//     for (unsigned t : adjacent(v)) { 
-//         addEdge(t, u);
-//         decrementDegree(t);
-//     }
-
-//     // 5. 检查合并后的节点 u 的状态，如果其度数变高，可能需要将其移到 spillWorklist
-//     if (!precolored.count(u)) {
-//         int K = isFPVReg(u) ? K_fp : K_int;
-//         if (degree.at(u) >= K && freezeWorklist.count(u)) {
-//             freezeWorklist.erase(u);
-//             spillWorklist.insert(u);
-//         }
-//     }
-// }
-
 void RISCv64RegAlloc::combine(unsigned u, unsigned v) {
     freezeWorklist.erase(v);
     spillWorklist.erase(v);
@@ -1377,9 +1380,9 @@ void RISCv64RegAlloc::applyColoring() {
                             // 使用 setPReg 将虚拟寄存器转换为物理寄存器
                             reg_op->setPReg(color_map.at(vreg));
                         } else {
-                            // 如果一个vreg在成功分配后仍然没有颜色，这是一个错误
-                            std::cerr << "FATAL: Virtual register %vreg" << vreg << " has no color after allocation!\n";
-                            assert(false && "Virtual register has no color after allocation!");
+                            // 如果一个vreg在成功分配后仍然没有颜色，可能是紧急溢出
+                            // std::cerr << "WARNING: Virtual register %vreg" << vreg << " has no color after allocation, treating as spilled\n";
+                            // 在紧急溢出情况下，使用临时寄存器
                             reg_op->setPReg(PhysicalReg::T6);
                         }
                     }
@@ -1391,7 +1394,7 @@ void RISCv64RegAlloc::applyColoring() {
                         if (color_map.count(vreg)) {
                             reg_op->setPReg(color_map.at(vreg));
                         } else {
-                            assert(false && "Virtual register in memory operand has no color!");
+                            // std::cerr << "WARNING: Virtual register in memory operand has no color, using T6\n";
                             reg_op->setPReg(PhysicalReg::T6);
                         }
                     }
@@ -1483,7 +1486,6 @@ std::string RISCv64RegAlloc::regToString(PhysicalReg reg) {
 std::string RISCv64RegAlloc::regIdToString(unsigned id) {
     const unsigned offset = static_cast<unsigned>(PhysicalReg::PHYS_REG_START_ID);
 
-    // 使用更健壮的检查方式
     if (id >= offset && precolored.count(id)) {
         // 先减去偏移量，得到原始的、小的枚举值
         PhysicalReg reg = static_cast<PhysicalReg>(id - offset);

src/include/backend/RISCv64/Optimize/DivStrengthReduction.h

@@ -0,0 +1,30 @@
+#ifndef RISCV64_DIV_STRENGTH_REDUCTION_H
+#define RISCV64_DIV_STRENGTH_REDUCTION_H
+
+#include "RISCv64LLIR.h"
+#include "Pass.h"
+
+namespace sysy {
+
+/**
+ * @class DivStrengthReduction
+ * @brief 除法强度削弱优化器
+ * * 将除法运算转换为乘法运算，使用magic number算法
+ * 适用于除数为常数的情况，可以显著提高性能
+ */
+class DivStrengthReduction : public Pass {
+public:
+    static char ID;
+    
+    DivStrengthReduction() : Pass("div-strength-reduction", Granularity::Function, PassKind::Optimization) {}
+    
+    void *getPassID() const override { return &ID; }
+    
+    bool runOnFunction(Function *F, AnalysisManager& AM) override;
+    
+    void runOnMachineFunction(MachineFunction* mfunc);
+};
+
+} // namespace sysy
+
+#endif // RISCV64_DIV_STRENGTH_REDUCTION_H

src/include/backend/RISCv64/RISCv64AsmPrinter.h

@@ -20,6 +20,8 @@ public:
     void setStream(std::ostream& os) { OS = &os; }
     // 辅助函数
     std::string regToString(PhysicalReg reg);
+    std::string formatInstr(const MachineInstr *instr);
+
 private:
     // 打印各个部分
     void printBasicBlock(MachineBasicBlock* mbb, bool debug = false);

src/include/backend/RISCv64/RISCv64Backend.h

@@ -24,7 +24,7 @@ private:
 
     // 私有辅助函数，用于根据类型计算其占用的字节数。
     unsigned getTypeSizeInBytes(Type* type);
-
+    
     Module* module;
 };
 

src/include/backend/RISCv64/RISCv64ISel.h

@@ -3,6 +3,12 @@
 
 #include "RISCv64LLIR.h"
 
+// Forward declarations
+namespace sysy {
+    class GlobalValue;
+    class Value;
+}
+
 extern int DEBUG;
 extern int DEEPDEBUG;
 
@@ -17,7 +23,8 @@ public:
     // 公开接口，以便后续模块（如RegAlloc）可以查询或创建vreg
     unsigned getVReg(Value* val);
     unsigned getNewVReg() { return vreg_counter++; }
-    unsigned getNewVReg(Type* type); 
+    unsigned getNewVReg(Type* type);
+    unsigned getVRegCounter() const; 
     // 获取 vreg_map 的公共接口
     const std::map<Value*, unsigned>& getVRegMap() const { return vreg_map; }
     const std::map<unsigned, Value*>& getVRegValueMap() const { return vreg_to_value_map; }

src/include/backend/RISCv64/RISCv64LLIR.h

@@ -39,14 +39,14 @@ enum class PhysicalReg {
 
     // 用于内部表示物理寄存器在干扰图中的节点ID（一个简单的特殊ID，确保不与vreg_counter冲突）
     // 假设 vreg_counter 不会达到这么大的值
-    PHYS_REG_START_ID = 100000, 
+    PHYS_REG_START_ID = 1000000, 
     PHYS_REG_END_ID = PHYS_REG_START_ID + 320, // 预留足够的空间
 };
 
 // RISC-V 指令操作码枚举
 enum class RVOpcodes {
     // 算术指令
-    ADD, ADDI, ADDW, ADDIW, SUB, SUBW, MUL, MULW, DIV, DIVW, REM, REMW,
+    ADD, ADDI, ADDW, ADDIW, SUB, SUBW, MUL, MULW, MULH, DIV, DIVW, REM, REMW,
     // 逻辑指令
     XOR, XORI, OR, ORI, AND, ANDI,
     // 移位指令
@@ -280,14 +280,15 @@ private:
 // 栈帧信息
 struct StackFrameInfo {
     int locals_size = 0; // 仅为AllocaInst分配的大小
+    int locals_end_offset = 0; // 记录局部变量分配结束后的偏移量(相对于s0，为负)
     int spill_size = 0; // 仅为溢出分配的大小
     int total_size = 0; // 总大小
     int callee_saved_size = 0; // 保存寄存器的大小
     std::map<unsigned, int> alloca_offsets; // <AllocaInst的vreg, 栈偏移>
     std::map<unsigned, int> spill_offsets;  // <溢出vreg, 栈偏移>
     std::set<PhysicalReg> used_callee_saved_regs; // 使用的保存寄存器
-    std::map<unsigned, PhysicalReg> vreg_to_preg_map; // [新增] RegAlloc最终的分配结果
-    std::vector<PhysicalReg> callee_saved_regs_to_store; // [新增] 已排序的、需要存取的被调用者保存寄存器
+    std::map<unsigned, PhysicalReg> vreg_to_preg_map; // RegAlloc最终的分配结果
+    std::vector<PhysicalReg> callee_saved_regs_to_store; // 已排序的、需要存取的被调用者保存寄存器
 };
 
 // 机器函数

src/include/backend/RISCv64/RISCv64Passes.h

@@ -10,6 +10,8 @@
 #include "PrologueEpilogueInsertion.h"
 #include "EliminateFrameIndices.h"
 #include "Pass.h"
+#include "DivStrengthReduction.h"
+
 
 namespace sysy {
 

src/include/midend/IR.h

@@ -20,6 +20,10 @@
 #include <algorithm>
 
 namespace sysy {
+
+// Global cleanup function to release all statically allocated IR objects
+void cleanupIRPools();
+
 /**
  * \defgroup type Types
  * @brief Sysy的类型系统
@@ -83,6 +87,7 @@ class Type {
   auto as() const -> std::enable_if_t<std::is_base_of_v<Type, T>, T *> {
     return dynamic_cast<T *>(const_cast<Type *>(this));
   }
+  virtual void print(std::ostream& os) const;
 };
 
 class PointerType : public Type {
@@ -94,6 +99,9 @@ class PointerType : public Type {
 
  public:
   static PointerType* get(Type *baseType);  ///< 获取指向baseType的Pointer类型
+  
+  // Cleanup method to release all cached pointer types (call at program exit)
+  static void cleanup();
 
  public:
   Type* getBaseType() const { return baseType; }  ///< 获取指向的类型
@@ -111,6 +119,9 @@ class FunctionType : public Type {
  public:
   /// 获取返回值类型为returnType， 形参类型列表为paramTypes的Function类型
   static FunctionType* get(Type *returnType, const std::vector<Type *> &paramTypes = {});
+  
+  // Cleanup method to release all cached function types (call at program exit)
+  static void cleanup();
 
  public:
   Type* getReturnType() const { return returnType; }          ///< 获取返回值类信息
@@ -123,6 +134,9 @@ class ArrayType : public Type {
   // elements：数组的元素类型 (例如，int[3] 的 elementType 是 int)
   // numElements：该维度的大小 (例如，int[3] 的 numElements 是 3)
   static ArrayType *get(Type *elementType, unsigned numElements);
+  
+  // Cleanup method to release all cached array types (call at program exit)
+  static void cleanup();
 
   Type *getElementType() const { return elementType; }
   unsigned getNumElements() const { return numElements; }
@@ -202,9 +216,11 @@ class Use {
 
  public:
   unsigned getIndex() const { return index; }   ///< 返回value在User操作数中的位置
+  void setIndex(int newIndex) { index = newIndex; }  ///< 设置value在User操作数中的位置
   User* getUser() const { return user; }       ///< 返回使用者
   Value* getValue() const { return value; }    ///< 返回被使用的值
   void setValue(Value *newValue) { value = newValue; }  ///< 将被使用的值设置为newValue
+  void print(std::ostream& os) const;
 };
 
 //! The base class of all value types
@@ -229,7 +245,15 @@ class Value {
   std::list<std::shared_ptr<Use>>& getUses() { return uses; }   ///< 获取使用关系列表
   void addUse(const std::shared_ptr<Use> &use) { uses.push_back(use); }  ///< 添加使用关系
   void replaceAllUsesWith(Value *value);  ///< 将原来使用该value的使用者全变为使用给定参数value并修改相应use关系
-  void removeUse(const std::shared_ptr<Use> &use) { uses.remove(use); }  ///< 删除使用关系use
+  void removeUse(const std::shared_ptr<Use> &use) { 
+    assert(use != nullptr && "Use cannot be null");
+    assert(use->getValue() == this && "Use being removed does NOT point to this Value!");
+    auto it = std::find(uses.begin(), uses.end(), use);
+    assert(it != uses.end() && "Use not found in Value's uses");
+    uses.remove(use); 
+  }  ///< 删除使用关系use
+  void removeAllUses();
+  virtual void print(std::ostream& os) const = 0;  ///< 输出值信息到输出流
 };
 
 /**
@@ -356,6 +380,9 @@ public:
 
   // Static factory method to get a canonical ConstantValue from the pool
   static ConstantValue* get(Type* type, ConstantValVariant val);
+  
+  // Cleanup method to release all cached constants (call at program exit)
+  static void cleanup();
 
   // Helper methods to access constant values with appropriate casting
   int getInt() const {
@@ -394,6 +421,7 @@ public:
 
   virtual bool isZero() const = 0;
   virtual bool isOne() const = 0;
+  void print(std::ostream& os) const = 0;
 };
 
 class ConstantInteger : public ConstantValue {
@@ -420,6 +448,7 @@ public:
 
   bool isZero() const override { return constVal == 0; }
   bool isOne() const override { return constVal == 1; }
+  void print(std::ostream& os) const;
 };
 
 class ConstantFloating : public ConstantValue {
@@ -446,6 +475,7 @@ public:
 
   bool isZero() const override { return constFVal == 0.0f; }
   bool isOne() const override { return constFVal == 1.0f; }
+  void print(std::ostream& os) const;
 };
 
 class UndefinedValue : public ConstantValue {
@@ -460,6 +490,9 @@ protected:
 
 public:
   static UndefinedValue* get(Type* type);
+  
+  // Cleanup method to release all cached undefined values (call at program exit)
+  static void cleanup();
 
   size_t hash() const override {
     return std::hash<Type*>{}(getType());
@@ -477,6 +510,7 @@ public:
 
   bool isZero() const override { return false; }
   bool isOne() const override { return false; }
+  void print(std::ostream& os) const;
 };
 
 // --- End of refactored ConstantValue and related classes ---
@@ -617,6 +651,11 @@ public:
     }
   }  ///< 移除指定位置的指令
   iterator moveInst(iterator sourcePos, iterator targetPos, BasicBlock *block);
+  
+  /// 清理基本块中的所有使用关系
+  void cleanup();
+  
+  void print(std::ostream& os) const; 
 };
 
 //! User is the abstract base type of `Value` types which use other `Value` as
@@ -633,21 +672,6 @@ class User : public Value {
   explicit User(Type *type, const std::string &name = "") : Value(type, name) {}
 
  public:
-  // ~User() override {
-  //   // 当 User 对象被销毁时（例如，LoadInst 或 StoreInst 被删除时），
-  //   // 它必须通知它所使用的所有 Value，将对应的 Use 关系从它们的 uses 列表中移除。
-  //   // 这样可以防止 Value 的 uses 列表中出现悬空的 Use 对象。
-  //   for (const auto &use_ptr : operands) {
-  //     // 确保 use_ptr 非空，并且其内部指向的 Value* 也非空
-  //     // (虽然通常情况下不会为空，但为了健壮性考虑)
-  //     if (use_ptr && use_ptr->getValue()) {
-  //       use_ptr->getValue()->removeUse(use_ptr);
-  //     }
-  //   }
-  //   // operands 向量本身是 std::vector<std::shared_ptr<Use>>，
-  //   // 在此析构函数结束后，operands 向量会被销毁，其内部的 shared_ptr 也会被释放，
-  //   // 如果 shared_ptr 引用计数降为0，Use 对象本身也会被销毁。
-  // }
   unsigned getNumOperands() const { return operands.size(); }  ///< 获取操作数数量
   auto operand_begin() const { return operands.begin(); }              ///< 返回操作数列表的开头迭代器
   auto operand_end() const { return operands.end(); }                  ///< 返回操作数列表的结尾迭代器
@@ -657,11 +681,7 @@ class User : public Value {
     operands.emplace_back(std::make_shared<Use>(operands.size(), this, value));
     value->addUse(operands.back());
   }  ///< 增加操作数
-  void removeOperand(unsigned index) {
-    auto value = getOperand(index);
-    value->removeUse(operands[index]);
-    operands.erase(operands.begin() + index);
-  }  ///< 移除操作数
+  void removeOperand(unsigned index);
   template <typename ContainerT>
   void addOperands(const ContainerT &newoperands) {
     for (auto value : newoperands) {
@@ -670,6 +690,9 @@ class User : public Value {
   }                                                   ///< 增加多个操作数
   void replaceOperand(unsigned index, Value *value);  ///< 替换操作数
   void setOperand(unsigned index, Value *value);      ///< 设置操作数
+  
+  /// 清理用户的所有操作数使用关系
+  void cleanup();
 };
 
 /*!
@@ -704,6 +727,7 @@ class Instruction : public User {
     kFCmpGE = 0x1UL << 20,
     kAnd = 0x1UL << 21,
     kOr = 0x1UL << 22,
+    // kXor = 0x1UL << 46,
     // Unary
     kNeg = 0x1UL << 23,
     kNot = 0x1UL << 24,
@@ -728,6 +752,10 @@ class Instruction : public User {
     kPhi = 0x1UL << 39,
     kBitItoF = 0x1UL << 40,
     kBitFtoI = 0x1UL << 41,
+    kSrl = 0x1UL << 42, // 逻辑右移
+    kSll = 0x1UL << 43, // 逻辑左移
+    kSra = 0x1UL << 44, // 算术右移
+    kMulh = 0x1UL << 45
   };
 
 protected:
@@ -745,57 +773,57 @@ public:
   std::string getKindString() const{
     switch (kind) {
       case kInvalid:
-        return "Invalid";
+        return "invalid";
       case kAdd:
-        return "Add";
+        return "add";
       case kSub:
-        return "Sub";
+        return "sub";
       case kMul:
-        return "Mul";
+        return "mul";
       case kDiv:
-        return "Div";
+        return "sdiv";
       case kRem:
-        return "Rem";
+        return "srem";
       case kICmpEQ:
-        return "ICmpEQ";
+        return "icmp eq";
       case kICmpNE:
-        return "ICmpNE";
+        return "icmp ne";
       case kICmpLT:
-        return "ICmpLT";
+        return "icmp slt";
       case kICmpGT:
-        return "ICmpGT";
+        return "icmp sgt";
       case kICmpLE:
-        return "ICmpLE";
+        return "icmp sle";
       case kICmpGE:
-        return "ICmpGE";
+        return "icmp sge";
       case kFAdd:
-        return "FAdd";
+        return "fadd";
       case kFSub:
-        return "FSub";
+        return "fsub";
       case kFMul:
-        return "FMul";
+        return "fmul";
       case kFDiv:
-        return "FDiv";
+        return "fdiv";
       case kFCmpEQ:
-        return "FCmpEQ";
+        return "fcmp oeq";
       case kFCmpNE:
-        return "FCmpNE";
+        return "fcmp one";
       case kFCmpLT:
-        return "FCmpLT";
+        return "fcmp olt";
       case kFCmpGT:
-        return "FCmpGT";
+        return "fcmp ogt";
       case kFCmpLE:
-        return "FCmpLE";
+        return "fcmp ole";
       case kFCmpGE:
-        return "FCmpGE";
+        return "fcmp oge";
       case kAnd:
-        return "And";
+        return "and";
       case kOr:
-        return "Or";
+        return "or";
       case kNeg:
-        return "Neg";
+        return "neg";
       case kNot:
-        return "Not";
+        return "not";
       case kFNeg:
         return "FNeg";
       case kFNot:
@@ -803,27 +831,41 @@ public:
       case kFtoI:
         return "FtoI";
       case kItoF:
-        return "IToF";
+        return "iToF";
       case kCall:
-        return "Call";
+        return "call";
       case kCondBr:
-        return "CondBr";
+        return "condBr";
       case kBr:
-        return "Br";
+        return "br";
       case kReturn:
-        return "Return";
+        return "return";
+      case kUnreachable:
+        return "unreachable";
       case kAlloca:
-        return "Alloca";
+        return "alloca";
       case kLoad:
-        return "Load";
+        return "load";
       case kStore:
-        return "Store";
+        return "store";
       case kGetElementPtr:
-        return "GetElementPtr";
+        return "getElementPtr";
       case kMemset:
-        return "Memset";
+        return "memset";
       case kPhi:
-        return "Phi";
+        return "phi";
+      case kBitItoF:
+        return "BitItoF";
+      case kBitFtoI:
+        return "BitFtoI";
+      case kSrl:
+        return "lshr";
+      case kSll:
+        return "shl";
+      case kSra:
+        return "ashr";
+      case kMulh:
+        return "mulh";
       default:
         return "Unknown";
     }
@@ -835,11 +877,15 @@ public:
 
   bool isBinary() const {
     static constexpr uint64_t BinaryOpMask =
-        (kAdd | kSub | kMul | kDiv | kRem | kAnd | kOr) |
-        (kICmpEQ | kICmpNE | kICmpLT | kICmpGT | kICmpLE | kICmpGE) |
+        (kAdd | kSub | kMul | kDiv | kRem | kAnd | kOr | kSra | kSrl | kSll | kMulh) |
+        (kICmpEQ | kICmpNE | kICmpLT | kICmpGT | kICmpLE | kICmpGE);
+    return kind & BinaryOpMask;
+  }
+  bool isFPBinary() const {
+    static constexpr uint64_t FPBinaryOpMask =
         (kFAdd | kFSub | kFMul | kFDiv) |
         (kFCmpEQ | kFCmpNE | kFCmpLT | kFCmpGT | kFCmpLE | kFCmpGE);
-    return kind & BinaryOpMask;
+    return kind & FPBinaryOpMask;
   }
   bool isUnary() const {
     static constexpr uint64_t UnaryOpMask = 
@@ -886,6 +932,10 @@ public:
     static constexpr uint64_t DefineOpMask = kAlloca | kStore | kPhi;
     return (kind & DefineOpMask) != 0U;
   } 
+  
+  virtual ~Instruction() = default;
+  
+  virtual void print(std::ostream& os) const = 0;
 }; // class Instruction
 
 class Function;
@@ -907,58 +957,62 @@ class PhiInst : public Instruction {
           const std::string &name = "")
       : Instruction(Kind::kPhi, type, parent, name), vsize(rhs.size()) {
     assert(rhs.size() == Blocks.size() && "PhiInst: rhs and Blocks must have the same size");
-    for(size_t i = 0; i < rhs.size(); ++i) {
+    for(size_t i = 0; i < vsize; ++i) {
       addOperand(rhs[i]);
+      addOperand(Blocks[i]);
       blk2val[Blocks[i]] = rhs[i];
     }
   }
 
  public:
-  Value* getValue(unsigned k) const {return getOperand(2 * k);}  ///< 获取位置为k的值
-  BasicBlock* getBlock(unsigned k) const {return dynamic_cast<BasicBlock*>(getOperand(2 * k + 1));}
-  //增加llvm同名方法实现获取value和block
-  Value* getIncomingValue(unsigned k) const {return getOperand(2 * k);}  ///< 获取位置为k的值
-  BasicBlock* getIncomingBlock(unsigned k) const {return dynamic_cast<BasicBlock*>(getOperand(2 * k + 1));}
-
-  Value* getIncomingValue(BasicBlock* blk) const {
-    return getvalfromBlk(blk);
-  }  ///< 获取指定基本块的传入值
-
-  BasicBlock* getIncomingBlock(Value* val) const {
-    return getBlkfromVal(val);
-  }  ///< 获取指定值的传入基本块
-
-  void replaceIncoming(BasicBlock *oldBlock, BasicBlock *newBlock, Value *newValue){
-    delBlk(oldBlock);
-    addIncoming(newValue, newBlock);
-  }
-
-  auto& getincomings() const {return blk2val;}  ///< 获取所有的基本块和对应的值
-
-  Value* getvalfromBlk(BasicBlock* blk) const ;
-  BasicBlock* getBlkfromVal(Value* val) const ;
-
   unsigned getNumIncomingValues() const { return vsize; }  ///< 获取传入值的数量
+  Value *getIncomingValue(unsigned Idx) const { return getOperand(Idx * 2); }  ///< 获取指定位置的传入值
+  BasicBlock *getIncomingBlock(unsigned Idx) const {return dynamic_cast<BasicBlock *>(getOperand(Idx * 2 + 1)); }  ///< 获取指定位置的传入基本块
+
+  Value* getValfromBlk(BasicBlock* block);
+  BasicBlock* getBlkfromVal(Value* value);
+  auto getIncomingValues() const {
+    std::vector<std::pair<BasicBlock*, Value*>> result;
+    for (const auto& [block, value] : blk2val) {
+      result.emplace_back(block, value);
+    }
+    return result;
+  }
   void addIncoming(Value *value, BasicBlock *block) {
-    assert(value && block && "PhiInst: value and block must not be null");
+    assert(value && block && "PhiInst: value and block cannot be null");
     addOperand(value);
     addOperand(block);
     blk2val[block] = value;
     vsize++;
   }  ///< 添加传入值和对应的基本块
-
-  void removeIncoming(BasicBlock *block){
-    delBlk(block);
-  }
-
-  void delValue(Value* val);
-  void delBlk(BasicBlock* blk);
-
-  void replaceBlk(BasicBlock* newBlk, unsigned k);
-  void replaceold2new(BasicBlock* oldBlk, BasicBlock* newBlk);
-  void refreshB2VMap();
-
+  void removeIncoming(unsigned Idx) {
+    assert(Idx < vsize && "PhiInst: Index out of bounds");
+    auto blk = getIncomingBlock(Idx);
+    removeOperand(Idx * 2 + 1);  // Remove block
+    removeOperand(Idx * 2);  // Remove value
+    blk2val.erase(blk);
+    vsize--;
+  }  ///< 移除指定位置的传入值和对应的基本块
+  // 移除指定的传入值或基本块
+  void removeIncomingValue(Value *value);
+  void removeIncomingBlock(BasicBlock *block);
+  // 设置指定位置的传入值或基本块
+  void setIncomingValue(unsigned Idx, Value *value);
+  void setIncomingBlock(unsigned Idx, BasicBlock *block);
+  // 替换指定位置的传入值或基本块（原理是删除再添加）保留旧块或者旧值
+  void replaceIncomingValue(Value *oldValue, Value *newValue);
+  void replaceIncomingBlock(BasicBlock *oldBlock, BasicBlock *newBlock);
+  // 替换指定位置的传入值或基本块（原理是删除再添加）
+  void replaceIncomingValue(Value *oldValue, Value *newValue, BasicBlock *newBlock);
+  void replaceIncomingBlock(BasicBlock *oldBlock, BasicBlock *newBlock, Value *newValue);
+  void refreshMap() {
+    blk2val.clear();
+    for (unsigned i = 0; i < vsize; ++i) {
+      blk2val[getIncomingBlock(i)] = getIncomingValue(i);
+    }
+  }  ///< 刷新块到值的映射关系
   auto getValues() { return make_range(std::next(operand_begin()), operand_end()); }
+  void print(std::ostream& os) const override;
 };
 
 
@@ -967,16 +1021,14 @@ class CallInst : public Instruction {
   friend class IRBuilder;
 
 protected:
-  CallInst(Function *callee, const std::vector<Value *> &args = {},
-           BasicBlock *parent = nullptr, const std::string &name = "");
-
+  CallInst(Function *callee, const std::vector<Value *> &args, BasicBlock *parent = nullptr, const std::string &name = "");
 
 public:
-  Function* getCallee() const;
+  Function *getCallee() const;
   auto getArguments() const {
     return make_range(std::next(operand_begin()), operand_end());
   }
-
+  void print(std::ostream& os) const override;
 }; // class CallInst
 
 //! Unary instruction, includes '!', '-' and type conversion.
@@ -994,7 +1046,7 @@ protected:
 
 public:
   Value* getOperand() const { return User::getOperand(0); }
-
+  void print(std::ostream& os) const override;
 }; // class UnaryInst
 
 //! Binary instruction, e.g., arithmatic, relation, logic, etc.
@@ -1073,6 +1125,7 @@ public:
         // 后端处理数组访存操作时需要创建计算地址的指令，需要在外部构造 BinaryInst 对象
         return new BinaryInst(kind, type, lhs, rhs, parent, name);
   }
+  void print(std::ostream& os) const override;
 }; // class BinaryInst
 
 //! The return statement
@@ -1093,6 +1146,7 @@ class ReturnInst : public Instruction {
   Value* getReturnValue() const {
     return hasReturnValue() ? getOperand(0) : nullptr;
   }
+  void print(std::ostream& os) const override;
 };
 
 //! Unconditional branch
@@ -1122,7 +1176,7 @@ public:
     }
     return succs;
   }
-
+  void print(std::ostream& os) const override;
 }; // class UncondBrInst
 
 //! Conditional branch
@@ -1162,7 +1216,7 @@ public:
     }
     return succs;
   }
-
+  void print(std::ostream& os) const override;
 }; // class CondBrInst
 
 class UnreachableInst : public Instruction {
@@ -1170,7 +1224,7 @@ public:
   // 构造函数：设置指令类型为 kUnreachable
   explicit UnreachableInst(const std::string& name, BasicBlock *parent = nullptr)
       : Instruction(kUnreachable, Type::getVoidType(), parent, "") {}
-
+  void print(std::ostream& os) const { os << "unreachable"; }
 };
 
 //! Allocate memory for stack variables, used for non-global variable declartion
@@ -1188,7 +1242,7 @@ public:
   Type* getAllocatedType() const {
     return getType()->as<PointerType>()->getBaseType();
   }  ///< 获取分配的类型
-
+  void print(std::ostream& os) const override;
 }; // class AllocaInst
 
 
@@ -1226,6 +1280,7 @@ public:
                                     BasicBlock *parent = nullptr, const std::string &name = "") {
     return new GetElementPtrInst(resultType, basePointer, indices, parent, name);
   }
+  void print(std::ostream& os) const override;
 };
 
 //! Load a value from memory address specified by a pointer value
@@ -1243,7 +1298,7 @@ protected:
 
 public:
   Value* getPointer() const { return getOperand(0); }
-  
+  void print(std::ostream& os) const override;
 }; // class LoadInst
 
 //! Store a value to memory address specified by a pointer value
@@ -1262,7 +1317,7 @@ protected:
 public:
   Value* getValue() const { return getOperand(0); }
   Value* getPointer() const { return getOperand(1); }
-
+  void print(std::ostream& os) const override;
 }; // class StoreInst
 
 //! Memset instruction
@@ -1292,7 +1347,7 @@ public:
   Value* getBegin() const { return getOperand(1); }
   Value* getSize() const { return getOperand(2); }
   Value* getValue() const { return getOperand(3); }
-
+  void print(std::ostream& os) const override;
 };
 
 class GlobalValue;
@@ -1310,6 +1365,11 @@ public:
 public:
   Function* getParent() const { return func; }
   int getIndex() const { return index; }
+  
+  /// 清理参数的使用关系
+  void cleanup();
+  
+  void print(std::ostream& os) const;
 };
 
 
@@ -1375,8 +1435,19 @@ protected:
     auto is_same_ptr = [blockToRemove](const std::unique_ptr<BasicBlock> &ptr) { return ptr.get() == blockToRemove; };
     blocks.remove_if(is_same_ptr);
   }
+  BasicBlock* addBasicBlock(const std::string &name, BasicBlock *before) {
+    // 在指定的基本块之前添加一个新的基本块
+    auto it = std::find_if(blocks.begin(), blocks.end(),
+                           [before](const std::unique_ptr<BasicBlock> &ptr) { return ptr.get() == before; });
+    if (it != blocks.end()) {
+      auto newblk = blocks.emplace(it, std::make_unique<BasicBlock>(this, name));
+      return newblk->get();  // 返回新添加的基本块指针
+    }
+    assert(false && "BasicBlock to insert before not found!"); 
+    return nullptr;  // 如果没有找到指定的基本块，则返回nullptr
+  }  ///< 添加一个新的基本块到某个基本块之前
   BasicBlock* addBasicBlock(const std::string &name = "") {
-    blocks.emplace_back(new BasicBlock(this, name));
+    blocks.emplace_back(std::make_unique<BasicBlock>(this, name));
     return blocks.back().get();
   }
   BasicBlock* addBasicBlock(BasicBlock *block) {
@@ -1387,6 +1458,11 @@ protected:
     blocks.emplace_front(block);
     return block;
   }
+  
+  /// 清理函数中的所有使用关系
+  void cleanup();
+  
+  void print(std::ostream& os) const;
 };
 
 //! Global value declared at file scope
@@ -1452,6 +1528,7 @@ public:
     return getByIndex(index);
   }  ///< 通过多维索引indices获取初始值
   const ValueCounter& getInitValues() const { return initValues; } 
+  void print(std::ostream& os) const;
 }; // class GlobalValue
 
 
@@ -1509,6 +1586,8 @@ class ConstantVariable : public Value {
     return getByIndex(index);
   }                                                        ///< 通过多维索引indices获取初始值
   const ValueCounter& getInitValues() const { return initValues; }                           ///< 获取初始值
+  void print(std::ostream& os) const;
+  void print_init(std::ostream& os) const;
 };
 
 using SymbolTableNode = struct SymbolTableNode {
@@ -1531,6 +1610,8 @@ class SymbolTable {
 
   Value* getVariable(const std::string &name) const;  ///< 根据名字name以及当前作用域获取变量
   Value* addVariable(const std::string &name, Value *variable);               ///< 添加变量
+  void registerParameterName(const std::string &name);                        ///< 注册函数参数名字，避免alloca重名
+  void addVariableDirectly(const std::string &name, Value *variable);        ///< 直接添加变量到当前作用域，不重命名
   std::vector<std::unique_ptr<GlobalValue>>& getGlobals();                  ///< 获取全局变量列表
   const std::vector<std::unique_ptr<ConstantVariable>>& getConsts() const;  ///< 获取全局常量列表
   void enterNewScope();                                                              ///< 进入新的作用域
@@ -1538,6 +1619,9 @@ class SymbolTable {
   bool isInGlobalScope() const;                                              ///< 是否位于全局作用域
   void enterGlobalScope();                                                           ///< 进入全局作用域
   bool isCurNodeNull() { return curNode == nullptr; }
+  
+  /// 清理符号表中的所有内容
+  void cleanup();
 };
 
 //! IR unit for representing a SysY compile unit
@@ -1590,6 +1674,12 @@ class Module {
   void addVariable(const std::string &name, AllocaInst *variable) {
     variableTable.addVariable(name, variable);
   }  ///< 添加变量
+  void addVariableDirectly(const std::string &name, AllocaInst *variable) {
+    variableTable.addVariableDirectly(name, variable);
+  }  ///< 直接添加变量到当前作用域，不重命名
+  void registerParameterName(const std::string &name) {
+    variableTable.registerParameterName(name);
+  }  ///< 注册函数参数名字，避免alloca重名
   Value* getVariable(const std::string &name) {
     return variableTable.getVariable(name);
   }  ///< 根据名字name和当前作用域获取变量
@@ -1602,7 +1692,7 @@ class Module {
   }  ///< 获取函数
   Function* getExternalFunction(const std::string &name) const {
     auto result = externalFunctions.find(name);
-    if (result == functions.end()) {
+    if (result == externalFunctions.end()) {
       return nullptr;
     }
     return result->second.get();
@@ -1622,6 +1712,11 @@ class Module {
   void leaveScope() { variableTable.leaveScope(); }  ///< 离开作用域
 
   bool isInGlobalArea() const { return variableTable.isInGlobalScope(); }  ///< 是否位于全局作用域
+
+  /// 清理模块中的所有对象，包括函数、基本块、指令等
+  void cleanup();
+
+  void print(std::ostream& os) const;
 };
 
 /*!

src/include/midend/IRBuilder.h

@@ -217,6 +217,18 @@ class IRBuilder {
   BinaryInst * createOrInst(Value *lhs, Value *rhs, const std::string &name = "") {
     return createBinaryInst(Instruction::kOr, Type::getIntType(), lhs, rhs, name);
   }  ///< 创建按位或指令
+  BinaryInst * createSllInst(Value *lhs, Value *rhs, const std::string &name = "") {
+    return createBinaryInst(Instruction::kSll, Type::getIntType(), lhs, rhs, name);
+  }  ///< 创建逻辑左移指令
+  BinaryInst * createSrlInst(Value *lhs, Value *rhs, const std::string &name = "") {
+    return createBinaryInst(Instruction::kSrl, Type::getIntType(), lhs, rhs, name);
+  }  ///< 创建逻辑右移指令
+  BinaryInst * createSraInst(Value *lhs, Value *rhs, const std::string &name = "") {
+    return createBinaryInst(Instruction::kSra, Type::getIntType(), lhs, rhs, name);
+  }  ///< 创建算术右移指令
+  BinaryInst * createMulhInst(Value *lhs, Value *rhs, const std::string &name = "") {
+    return createBinaryInst(Instruction::kMulh, Type::getIntType(), lhs, rhs, name);
+  }  ///< 创建高位乘法指令
   CallInst * createCallInst(Function *callee, const std::vector<Value *> &args, const std::string &name = "") {
     std::string newName;
     if (name.empty() && callee->getReturnType() != Type::getVoidType()) {
@@ -344,38 +356,31 @@ class IRBuilder {
     Type *currentWalkType = pointerType->as<PointerType>()->getBaseType();
 
     // 遍历所有索引来深入类型层次结构。
-    // `indices` 向量包含了所有 GEP 索引，包括由 `visitLValue` 等函数添加的初始 `0` 索引。
+    // 重要：第一个索引总是用于"解引用"指针，后续索引才用于数组/结构体的索引
     for (int i = 0; i < indices.size(); ++i) {
-        if (currentWalkType->isArray()) {
-            // 情况一：当前遍历类型是 `ArrayType`。
-            // 索引用于选择数组元素，`currentWalkType` 更新为数组的元素类型。
-            currentWalkType = currentWalkType->as<ArrayType>()->getElementType();
-        } else if (currentWalkType->isPointer()) {
-            // 情况二：当前遍历类型是 `PointerType`。
-            // 这意味着我们正在通过一个指针来访问其指向的内存。
-            // 索引用于选择该指针所指向的“数组”的元素。
-            // `currentWalkType` 更新为该指针所指向的基础类型。
-            // 例如：如果 `currentWalkType` 是 `i32*`，它将变为 `i32`。
-            // 如果 `currentWalkType` 是 `[10 x i32]*`，它将变为 `[10 x i32]`。
-            currentWalkType = currentWalkType->as<PointerType>()->getBaseType();
+        if (i == 0) {
+            // 第一个索引：总是用于"解引用"基指针，不改变currentWalkType
+            // 例如：对于 `[4 x i32]* ptr, i32 0`，第一个0只是说"访问ptr指向的对象"
+            // currentWalkType 保持为 `[4 x i32]`
+            continue;
         } else {
-            // 情况三：当前遍历类型是标量类型 (例如 `i32`, `float` 等非聚合、非指针类型)。
-            //
-            // 如果 `currentWalkType` 是标量，并且当前索引 `i` **不是** `indices` 向量中的最后一个索引，
-            // 这意味着尝试对一个标量类型进行进一步的结构性索引，这是**无效的**。
-            // 例如：`int x; x[0];` 对应的 GEP 链中，`x` 的类型是 `i32`，再加 `[0]` 索引就是错误。
-            //
-            // 如果 `currentWalkType` 是标量，且这是**最后一个索引** (`i == indices.size() - 1`)，
-            // 那么 GEP 是合法的，它只是计算一个偏移地址，最终的类型就是这个标量类型。
-            // 此时 `currentWalkType` 保持不变，循环结束。
-            if (i < indices.size() - 1) { 
-                assert(false && "Invalid GEP indexing: attempting to index into a non-aggregate/non-pointer type with further indices.");
-                return nullptr; // 返回空指针表示类型推断失败
+            // 后续索引：用于实际的数组/结构体索引
+            if (currentWalkType->isArray()) {
+                // 数组索引：选择数组中的元素
+                currentWalkType = currentWalkType->as<ArrayType>()->getElementType();
+            } else if (currentWalkType->isPointer()) {
+                // 指针索引：解引用指针并继续
+                currentWalkType = currentWalkType->as<PointerType>()->getBaseType();
+            } else {
+                // 标量类型：不能进一步索引
+                if (i < indices.size() - 1) { 
+                    assert(false && "Invalid GEP indexing: attempting to index into a non-aggregate/non-pointer type with further indices.");
+                    return nullptr;
+                }
             }
-            // 如果是最后一个索引，且当前类型是标量，则类型保持不变，这是合法的。
-            // 循环会自然结束，返回正确的 `currentWalkType`。
         }
     }
+    
     // 所有索引处理完毕后，`currentWalkType` 就是 GEP 指令最终计算出的地址所指向的元素的类型。
     return currentWalkType;
   }

src/include/midend/Pass/Analysis/AliasAnalysis.h

@@ -0,0 +1,246 @@
+#pragma once
+
+#include "IR.h"
+#include "Pass.h"
+#include <map>
+#include <set>
+#include <vector>
+#include <memory>
+
+namespace sysy {
+
+// 前向声明
+class MemoryLocation;
+class AliasAnalysisResult;
+
+/**
+ * @brief 别名关系类型
+ * 按风险等级递增排序
+ */
+enum class AliasType {
+  NO_ALIAS = 0,        // 确定无别名 (不同的局部数组)
+  SELF_ALIAS = 1,      // 自别名 (同一数组的不同索引)
+  POSSIBLE_ALIAS = 2,  // 可能有别名 (函数参数数组)
+  UNKNOWN_ALIAS = 3    // 未知 (保守估计)
+};
+
+/**
+ * @brief 内存位置信息
+ * 描述一个内存访问的基础信息
+ */
+struct MemoryLocation {
+  Value* basePointer;              // 基指针 (剥离GEP后的真实基址)
+  Value* accessPointer;            // 访问指针 (包含索引信息)
+  
+  // 分类信息
+  bool isLocalArray;               // 是否为局部数组
+  bool isFunctionParameter;        // 是否为函数参数
+  bool isGlobalArray;              // 是否为全局数组
+  
+  // 索引信息
+  std::vector<Value*> indices;     // GEP索引列表
+  bool hasConstantIndices;         // 是否为常量索引
+  bool hasLoopVariableIndex;       // 是否包含循环变量
+  int constantOffset;              // 常量偏移量 (仅当全部为常量时有效)
+  
+  // 访问模式
+  bool hasReads;                   // 是否有读操作
+  bool hasWrites;                  // 是否有写操作
+  std::vector<Instruction*> accessInsts; // 所有访问指令
+  
+  MemoryLocation(Value* base, Value* access) 
+    : basePointer(base), accessPointer(access), 
+      isLocalArray(false), isFunctionParameter(false), isGlobalArray(false),
+      hasConstantIndices(false), hasLoopVariableIndex(false), constantOffset(0),
+      hasReads(false), hasWrites(false) {}
+};
+
+/**
+ * @brief 别名分析结果
+ * 存储一个函数的完整别名分析信息
+ */
+class AliasAnalysisResult : public AnalysisResultBase {
+public:
+  AliasAnalysisResult(Function *F) : AssociatedFunction(F) {}
+  ~AliasAnalysisResult() override = default;
+
+  // ========== 基础查询接口 ==========
+  
+  /**
+   * 查询两个指针之间的别名关系
+   */
+  AliasType queryAlias(Value* ptr1, Value* ptr2) const;
+  
+  /**
+   * 查询指针的内存位置信息
+   */
+  const MemoryLocation* getMemoryLocation(Value* ptr) const;
+  
+  /**
+   * 获取所有内存位置
+   */
+  const std::map<Value*, std::unique_ptr<MemoryLocation>>& getAllMemoryLocations() const {
+    return LocationMap;
+  }
+  
+  // ========== 高级查询接口 ==========
+  
+  /**
+   * 检查指针是否为局部数组
+   */
+  bool isLocalArray(Value* ptr) const;
+  
+  /**
+   * 检查指针是否为函数参数数组
+   */
+  bool isFunctionParameter(Value* ptr) const;
+  
+  /**
+   * 检查指针是否为全局数组
+   */
+  bool isGlobalArray(Value* ptr) const;
+  
+  /**
+   * 检查指针是否使用常量索引
+   */
+  bool hasConstantAccess(Value* ptr) const;
+  
+  // ========== 统计接口 ==========
+  
+  /**
+   * 获取各类别名类型的统计信息
+   */
+  struct Statistics {
+    int totalQueries;
+    int noAlias;
+    int selfAlias;
+    int possibleAlias;
+    int unknownAlias;
+    int localArrays;
+    int functionParameters;
+    int globalArrays;
+    int constantAccesses;
+  };
+  
+  Statistics getStatistics() const;
+  
+  /**
+   * 打印别名分析结果 (调试用)
+   */
+  void print() const;
+  void printStatics() const;
+  // ========== 内部方法 ==========
+  
+  void addMemoryLocation(std::unique_ptr<MemoryLocation> location);
+  void addAliasRelation(Value* ptr1, Value* ptr2, AliasType type);
+  
+  // ========== 公开数据成员 (供Pass使用) ==========
+  std::map<Value*, std::unique_ptr<MemoryLocation>> LocationMap;  // 内存位置映射
+  std::map<std::pair<Value*, Value*>, AliasType> AliasMap;        // 别名关系缓存
+
+private:
+  Function *AssociatedFunction;                                    // 关联的函数
+  
+  // 分类存储
+  std::vector<Argument*> ArrayParameters;                         // 数组参数
+  std::vector<AllocaInst*> LocalArrays;                          // 局部数组
+  std::set<GlobalValue*> AccessedGlobals;                        // 访问的全局变量
+};
+
+/**
+ * @brief SysY语言特化的别名分析Pass
+ * 针对SysY语言特性优化的别名分析实现
+ */
+class SysYAliasAnalysisPass : public AnalysisPass {
+public:
+  // 唯一的 Pass ID
+  static void *ID;
+  // 在这里开启激进分析策略
+  SysYAliasAnalysisPass() : AnalysisPass("SysYAliasAnalysis", Pass::Granularity::Function), 
+                           aggressiveParameterMode(false), parameterOptimizationEnabled(false) {}
+
+  // 实现 getPassID
+  void *getPassID() const override { return &ID; }
+
+  // 核心运行方法
+  bool runOnFunction(Function *F, AnalysisManager &AM) override;
+
+  // 获取分析结果
+  std::unique_ptr<AnalysisResultBase> getResult() override { return std::move(CurrentResult); }
+  
+  // ========== 配置接口 ==========
+  
+  /**
+   * 启用针对SysY评测的激进优化模式
+   * 在这种模式下，假设不同参数不会传入相同数组
+   */
+  void enableSysYTestingMode() {
+    aggressiveParameterMode = true;
+    parameterOptimizationEnabled = true;
+  }
+  
+  /**
+   * 使用保守的默认模式（适合通用场景）
+   */
+  void useConservativeMode() {
+    aggressiveParameterMode = false;
+    parameterOptimizationEnabled = false;
+  }
+
+private:
+  std::unique_ptr<AliasAnalysisResult> CurrentResult; // 当前函数的分析结果
+  
+  // ========== 主要分析流程 ==========
+  
+  void collectMemoryAccesses(Function* F);              // 收集内存访问
+  void buildAliasRelations(Function* F);                // 构建别名关系
+  void optimizeForSysY(Function* F);                    // SysY特化优化
+  
+  // ========== 内存位置分析 ==========
+  
+  std::unique_ptr<MemoryLocation> createMemoryLocation(Value* ptr);
+  Value* getBasePointer(Value* ptr);                    // 获取基指针
+  void analyzeMemoryType(MemoryLocation* location);     // 分析内存类型
+  void analyzeIndexPattern(MemoryLocation* location);   // 分析索引模式
+  
+  // ========== 别名关系推断 ==========
+  
+  AliasType analyzeAliasBetween(MemoryLocation* loc1, MemoryLocation* loc2);
+  AliasType compareIndices(MemoryLocation* loc1, MemoryLocation* loc2);
+  AliasType compareLocalArrays(MemoryLocation* loc1, MemoryLocation* loc2);
+  AliasType compareParameters(MemoryLocation* loc1, MemoryLocation* loc2);
+  AliasType compareWithGlobal(MemoryLocation* loc1, MemoryLocation* loc2);
+  AliasType compareMixedTypes(MemoryLocation* loc1, MemoryLocation* loc2);
+  
+  // ========== SysY特化优化 ==========
+  
+  void applySysYConstraints(Function* F);               // 应用SysY语言约束
+  void optimizeParameterAnalysis(Function* F);          // 优化参数分析
+  void optimizeArrayAccessAnalysis(Function* F);        // 优化数组访问分析
+  
+  // ========== 配置和策略控制 ==========
+  
+  bool useAggressiveParameterAnalysis() const { return aggressiveParameterMode; }
+  bool enableParameterOptimization() const { return parameterOptimizationEnabled; }
+  void setAggressiveParameterMode(bool enable) { aggressiveParameterMode = enable; }
+  void setParameterOptimizationEnabled(bool enable) { parameterOptimizationEnabled = enable; }
+  
+  // ========== 辅助优化方法 ==========
+  
+  void optimizeConstantIndexAccesses();                 // 优化常量索引访问
+  void optimizeSequentialAccesses();                   // 优化顺序访问
+  
+  // ========== 辅助方法 ==========
+  
+  bool isConstantValue(Value* val);                     // 是否为常量
+  bool hasLoopVariableInIndices(const std::vector<Value*>& indices, Function* F);
+  int calculateConstantOffset(const std::vector<Value*>& indices);
+  void printStatistics() const;                         // 打印统计信息
+
+private:
+  // ========== 配置选项 ==========
+  bool aggressiveParameterMode = false;                // 激进的参数别名分析模式
+  bool parameterOptimizationEnabled = false;           // 启用参数优化
+};
+
+} // namespace sysy

src/include/midend/Pass/Analysis/CallGraphAnalysis.h

@@ -0,0 +1,242 @@
+#pragma once
+
+#include "IR.h"
+#include "Pass.h"
+#include <map>
+#include <set>
+#include <vector>
+#include <memory>
+#include <algorithm>
+#include <unordered_set>
+
+namespace sysy {
+
+// 前向声明
+class CallGraphAnalysisResult;
+
+/**
+ * @brief 调用图节点信息
+ * 存储单个函数在调用图中的信息
+ */
+struct CallGraphNode {
+    Function* function;                    // 关联的函数
+    std::set<Function*> callers;           // 调用此函数的函数集合
+    std::set<Function*> callees;           // 此函数调用的函数集合
+    
+    // 递归信息
+    bool isRecursive;                      // 是否参与递归调用
+    bool isSelfRecursive;                  // 是否自递归
+    int recursiveDepth;                    // 递归深度(-1表示无限递归)
+    
+    // 调用统计
+    size_t totalCallers;                   // 调用者总数
+    size_t totalCallees;                   // 被调用函数总数
+    size_t callSiteCount;                  // 调用点总数
+    
+    CallGraphNode(Function* f) : function(f), isRecursive(false), 
+        isSelfRecursive(false), recursiveDepth(0), totalCallers(0), 
+        totalCallees(0), callSiteCount(0) {}
+};
+
+/**
+ * @brief 调用图分析结果类
+ * 包含整个模块的调用图信息和查询接口
+ */
+class CallGraphAnalysisResult : public AnalysisResultBase {
+public:
+    CallGraphAnalysisResult(Module* M) : AssociatedModule(M) {}
+    ~CallGraphAnalysisResult() override = default;
+
+    // ========== 基础查询接口 ==========
+    
+    /**
+     * 获取函数的调用图节点
+     */
+    const CallGraphNode* getNode(Function* F) const {
+        auto it = nodes.find(F);
+        return (it != nodes.end()) ? it->second.get() : nullptr;
+    }
+    
+    /**
+     * 获取函数的调用图节点（非const版本）
+     */
+    CallGraphNode* getMutableNode(Function* F) {
+        auto it = nodes.find(F);
+        return (it != nodes.end()) ? it->second.get() : nullptr;
+    }
+    
+    /**
+     * 获取所有函数节点
+     */
+    const std::map<Function*, std::unique_ptr<CallGraphNode>>& getAllNodes() const {
+        return nodes;
+    }
+    
+    /**
+     * 检查函数是否存在于调用图中
+     */
+    bool hasFunction(Function* F) const {
+        return nodes.find(F) != nodes.end();
+    }
+
+    // ========== 调用关系查询 ==========
+    
+    /**
+     * 检查是否存在从caller到callee的调用
+     */
+    bool hasCallEdge(Function* caller, Function* callee) const {
+        auto node = getNode(caller);
+        return node && node->callees.count(callee) > 0;
+    }
+    
+    /**
+     * 获取函数的所有调用者
+     */
+    std::vector<Function*> getCallers(Function* F) const {
+        auto node = getNode(F);
+        if (!node) return {};
+        return std::vector<Function*>(node->callers.begin(), node->callers.end());
+    }
+    
+    /**
+     * 获取函数的所有被调用函数
+     */
+    std::vector<Function*> getCallees(Function* F) const {
+        auto node = getNode(F);
+        if (!node) return {};
+        return std::vector<Function*>(node->callees.begin(), node->callees.end());
+    }
+
+    // ========== 递归分析查询 ==========
+    
+    /**
+     * 检查函数是否参与递归调用
+     */
+    bool isRecursive(Function* F) const {
+        auto node = getNode(F);
+        return node && node->isRecursive;
+    }
+    
+    /**
+     * 检查函数是否自递归
+     */
+    bool isSelfRecursive(Function* F) const {
+        auto node = getNode(F);
+        return node && node->isSelfRecursive;
+    }
+    
+    /**
+     * 获取递归深度
+     */
+    int getRecursiveDepth(Function* F) const {
+        auto node = getNode(F);
+        return node ? node->recursiveDepth : 0;
+    }
+
+    // ========== 拓扑排序和SCC ==========
+    
+    /**
+     * 获取函数的拓扑排序结果
+     * 保证被调用函数在调用函数之前
+     */
+    const std::vector<Function*>& getTopologicalOrder() const {
+        return topologicalOrder;
+    }
+    
+    /**
+     * 获取强连通分量列表
+     * 每个SCC表示一个递归函数群
+     */
+    const std::vector<std::vector<Function*>>& getStronglyConnectedComponents() const {
+        return sccs;
+    }
+    
+    /**
+     * 获取函数所在的SCC索引
+     */
+    int getSCCIndex(Function* F) const {
+        auto it = functionToSCC.find(F);
+        return (it != functionToSCC.end()) ? it->second : -1;
+    }
+
+    // ========== 统计信息 ==========
+    
+    struct Statistics {
+        size_t totalFunctions;
+        size_t totalCallEdges;
+        size_t recursiveFunctions;
+        size_t selfRecursiveFunctions;
+        size_t stronglyConnectedComponents;
+        size_t maxSCCSize;
+        double avgCallersPerFunction;
+        double avgCalleesPerFunction;
+    };
+    
+    Statistics getStatistics() const;
+    
+    /**
+     * 打印调用图分析结果
+     */
+    void print() const;
+
+    // ========== 内部构建接口 ==========
+    
+    void addNode(Function* F);
+    void addCallEdge(Function* caller, Function* callee);
+    void computeTopologicalOrder();
+    void computeStronglyConnectedComponents();
+    void analyzeRecursion();
+
+private:
+    Module* AssociatedModule;                                      // 关联的模块
+    std::map<Function*, std::unique_ptr<CallGraphNode>> nodes;     // 调用图节点
+    std::vector<Function*> topologicalOrder;                      // 拓扑排序结果
+    std::vector<std::vector<Function*>> sccs;                     // 强连通分量
+    std::map<Function*, int> functionToSCC;                       // 函数到SCC的映射
+    
+    // 内部辅助方法
+    void dfsTopological(Function* F, std::unordered_set<Function*>& visited, 
+                       std::vector<Function*>& result);
+    void tarjanSCC();
+    void tarjanDFS(Function* F, int& index, std::vector<int>& indices, 
+                  std::vector<int>& lowlinks, std::vector<Function*>& stack, 
+                  std::unordered_set<Function*>& onStack);
+};
+
+/**
+ * @brief SysY调用图分析Pass
+ * Module级别的分析Pass，构建整个模块的函数调用图
+ */
+class CallGraphAnalysisPass : public AnalysisPass {
+public:
+    // 唯一的 Pass ID
+    static void* ID;
+
+    CallGraphAnalysisPass() : AnalysisPass("CallGraphAnalysis", Pass::Granularity::Module) {}
+
+    // 实现 getPassID
+    void* getPassID() const override { return &ID; }
+
+    // 核心运行方法
+    bool runOnModule(Module* M, AnalysisManager& AM) override;
+
+    // 获取分析结果
+    std::unique_ptr<AnalysisResultBase> getResult() override { return std::move(CurrentResult); }
+
+private:
+    std::unique_ptr<CallGraphAnalysisResult> CurrentResult;  // 当前模块的分析结果
+    
+    // ========== 主要分析流程 ==========
+    
+    void buildCallGraph(Module* M);                          // 构建调用图
+    void scanFunctionCalls(Function* F);                     // 扫描函数的调用
+    void processCallInstruction(CallInst* call, Function* caller);  // 处理调用指令
+    
+    // ========== 辅助方法 ==========
+    
+    bool isLibraryFunction(Function* F) const;               // 判断是否为标准库函数
+    bool isIntrinsicFunction(Function* F) const;             // 判断是否为内置函数
+    void printStatistics() const;                            // 打印统计信息
+};
+
+} // namespace sysy

src/include/midend/Pass/Analysis/Dom.h

@@ -6,30 +6,82 @@
 #include <set>
 #include <vector>
 #include <algorithm>
+#include <functional>
 
 namespace sysy {
 
-// 支配树分析结果类 (保持不变)
+// 支配树分析结果类
 class DominatorTree : public AnalysisResultBase {
 public:
     DominatorTree(Function* F);
+    // 获取指定基本块的所有支配者
     const std::set<BasicBlock*>* getDominators(BasicBlock* BB) const;
-    BasicBlock* getImmediateDominator(BasicBlock* BB) const;
-    const std::set<BasicBlock*>* getDominanceFrontier(BasicBlock* BB) const;
+    // 获取指定基本块的即时支配者 (Immediate Dominator)
+    BasicBlock* getImmediateDominator(BasicBlock* BB) const;  
+    // 获取指定基本块的支配边界 (Dominance Frontier)
+    const std::set<BasicBlock*>* getDominanceFrontier(BasicBlock* BB) const;   
+    // 获取指定基本块在支配树中的子节点
     const std::set<BasicBlock*>* getDominatorTreeChildren(BasicBlock* BB) const;
+    // 额外的 Getter：获取所有支配者、即时支配者和支配边界的完整映射（可选，主要用于调试或特定场景）
     const std::map<BasicBlock*, std::set<BasicBlock*>>& getDominatorsMap() const { return Dominators; }
     const std::map<BasicBlock*, BasicBlock*>& getIDomsMap() const { return IDoms; }
     const std::map<BasicBlock*, std::set<BasicBlock*>>& getDominanceFrontiersMap() const { return DominanceFrontiers; }
+
+    // 计算所有基本块的支配者集合
     void computeDominators(Function* F);
-    void computeIDoms(Function* F);
+    // 计算所有基本块的即时支配者（内部使用 Lengauer-Tarjan 算法）
+    void computeIDoms(Function* F); 
+    // 计算所有基本块的支配边界
     void computeDominanceFrontiers(Function* F);
+    // 计算支配树的结构（即每个节点的直接子节点）
     void computeDominatorTreeChildren(Function* F);
 private:
+    // 与该支配树关联的函数
     Function* AssociatedFunction;
-    std::map<BasicBlock*, std::set<BasicBlock*>> Dominators;
-    std::map<BasicBlock*, BasicBlock*> IDoms;
-    std::map<BasicBlock*, std::set<BasicBlock*>> DominanceFrontiers;
-    std::map<BasicBlock*, std::set<BasicBlock*>> DominatorTreeChildren;
+    std::map<BasicBlock*, std::set<BasicBlock*>> Dominators;       // 每个基本块的支配者集合
+    std::map<BasicBlock*, BasicBlock*> IDoms;                      // 每个基本块的即时支配者
+    std::map<BasicBlock*, std::set<BasicBlock*>> DominanceFrontiers; // 每个基本块的支配边界
+    std::map<BasicBlock*, std::set<BasicBlock*>> DominatorTreeChildren; // 支配树中每个基本块的子节点
+
+    // ==========================================================
+    // Lengauer-Tarjan 算法内部所需的数据结构和辅助函数
+    // 这些成员是私有的，以封装 LT 算法的复杂性并避免命名空间污染
+    // ==========================================================
+
+    // DFS 遍历相关：
+    std::map<BasicBlock*, int> dfnum_map;            // 存储每个基本块的 DFS 编号
+    std::vector<BasicBlock*> vertex_vec;             // 通过 DFS 编号反向查找对应的基本块指针
+    std::map<BasicBlock*, BasicBlock*> parent_map;   // 存储 DFS 树中每个基本块的父节点
+    int df_counter;                                  // DFS 计数器，也代表 DFS 遍历的总节点数 (N)
+
+    // 半支配者 (Semi-dominator) 相关：
+    std::map<BasicBlock*, BasicBlock*> sdom_map;     // 存储每个基本块的半支配者
+    std::map<BasicBlock*, BasicBlock*> idom_map;     // 存储每个基本块的即时支配者 (IDom)
+    std::map<BasicBlock*, std::vector<BasicBlock*>> bucket_map; // 桶结构，用于存储具有相同半支配者的节点，以延迟 IDom 计算
+
+    // 并查集 (Union-Find) 相关（用于 evalAndCompress 函数）：
+    std::map<BasicBlock*, BasicBlock*> ancestor_map; // 并查集中的父节点（用于路径压缩）
+    std::map<BasicBlock*, BasicBlock*> label_map;    // 并查集中，每个集合的代表节点（或其路径上 sdom 最小的节点）
+
+    // ==========================================================
+    // 辅助计算函数 (私有)
+    // ==========================================================
+
+    // 计算基本块的逆后序遍历 (Reverse Post Order, RPO) 顺序
+    // RPO 用于优化支配者计算和 LT 算法的效率
+    std::vector<BasicBlock*> computeReversePostOrder(Function* F);
+
+    // Lengauer-Tarjan 算法特定的辅助 DFS 函数
+    // 用于初始化 dfnum_map, vertex_vec, parent_map
+    void dfs_lt_helper(BasicBlock* u);                 
+    
+    // 结合了并查集的 Find 操作和 LT 算法的 Eval 操作
+    // 用于在路径压缩时更新 label，找到路径上 sdom 最小的节点
+    BasicBlock* evalAndCompress_lt_helper(BasicBlock* i); 
+    
+    // 并查集的 Link 操作
+    // 将 v_child 挂载到 u_parent 的并查集树下
+    void link_lt_helper(BasicBlock* u_parent, BasicBlock* v_child); 
 };
 
 

src/include/midend/Pass/Analysis/Loop.h

@@ -0,0 +1,618 @@
+#pragma once
+
+#include "Dom.h"
+#include "IR.h"
+#include "Pass.h"
+#include <algorithm>
+#include <functional>
+#include <map>
+#include <memory>
+#include <optional>
+#include <queue>
+#include <set>
+#include <vector>
+
+namespace sysy {
+
+// 前向声明
+class LoopAnalysisResult;
+class AliasAnalysisResult;
+class SideEffectAnalysisResult;
+
+/**
+ * @brief 表示一个识别出的循环。
+ */
+class Loop {
+private:
+  static int NextLoopID; // 静态变量用于分配唯一ID
+  int LoopID;
+public:
+  // 构造函数：指定循环头
+  Loop(BasicBlock *header) : Header(header), LoopID(NextLoopID++) {}
+
+  // 获取循环头
+  BasicBlock *getHeader() const { return Header; }
+
+  // 获取循环的名称 （基于ID）
+  std::string getName() const { return "loop_" + std::to_string(LoopID); }
+  // 获取循环体包含的所有基本块
+  const std::set<BasicBlock *> &getBlocks() const { return LoopBlocks; }
+
+  // 获取循环的出口基本块（即从循环内部跳转到循环外部的基本块）
+  const std::set<BasicBlock *> &getExitBlocks() const { return ExitBlocks; }
+
+  // 获取循环前置块（如果存在），可以为 nullptr
+  BasicBlock *getPreHeader() const { return PreHeader; }
+
+  // 获取直接包含此循环的父循环（如果存在），可以为 nullptr
+  Loop *getParentLoop() const { return ParentLoop; }
+
+  // 获取直接嵌套在此循环内的子循环
+  const std::vector<Loop *> &getNestedLoops() const { return NestedLoops; }
+
+  // 获取循环的层级 (0 表示最外层循环，1 表示嵌套一层，以此类推)
+  int getLoopLevel() const { return Level; }
+
+  // 检查一个基本块是否属于当前循环
+  bool contains(BasicBlock *BB) const { return LoopBlocks.count(BB); }
+
+  // 判断当前循环是否是最内层循环 (没有嵌套子循环)
+  bool isInnermost() const { return NestedLoops.empty(); }
+
+  // 获取循环的深度（从最外层开始计算）
+  int getLoopDepth() const { return Level + 1; }
+
+  // 获取循环体的大小（基本块数量）
+  size_t getLoopSize() const { return LoopBlocks.size(); }
+
+  // 检查循环是否有唯一的外部前驱（即是否有前置块）
+  bool hasUniquePreHeader() const { return PreHeader != nullptr; }
+
+  // 检查循环是否是最外层循环（没有父循环）
+  bool isOutermost() const { return getParentLoop() == nullptr; }
+
+  // 获取循环的所有出口（从循环内到循环外的基本块）
+  std::vector<BasicBlock*> getExitingBlocks() const {
+    std::vector<BasicBlock*> exitingBlocks;
+    for (BasicBlock* bb : LoopBlocks) {
+      for (BasicBlock* succ : bb->getSuccessors()) {
+        if (!contains(succ)) {
+          exitingBlocks.push_back(bb);
+          break; // 每个基本块只添加一次
+        }
+      }
+    }
+    return exitingBlocks;
+  }
+
+  // 判断循环是否是简单循环（只有一个回边）
+  bool isSimpleLoop() const {
+    int backEdgeCount = 0;
+    for (BasicBlock* pred : Header->getPredecessors()) {
+      if (contains(pred)) {
+        backEdgeCount++;
+      }
+    }
+    return backEdgeCount == 1;
+  }
+
+  /**
+   * 获取所有出口目标块 (循环外接收循环出口边的块)
+   * 使用场景: 循环后置处理、phi节点分析
+   */
+  std::vector<BasicBlock*> getExitTargetBlocks() const {
+    std::set<BasicBlock*> exitTargetSet;
+    for (BasicBlock* bb : LoopBlocks) {
+      for (BasicBlock* succ : bb->getSuccessors()) {
+        if (!contains(succ)) {
+          exitTargetSet.insert(succ);
+        }
+      }
+    }
+    return std::vector<BasicBlock*>(exitTargetSet.begin(), exitTargetSet.end());
+  }
+
+  /**
+   * 计算循环的"深度"相对于指定的祖先循环
+   * 使用场景: 相对深度计算、嵌套分析
+   */
+  int getRelativeDepth(Loop* ancestor) const {
+    if (this == ancestor) return 0;
+    
+    int depth = 0;
+    Loop* current = this->ParentLoop;
+    while (current && current != ancestor) {
+      depth++;
+      current = current->ParentLoop;
+    }
+    
+    return current == ancestor ? depth : -1; // -1表示不是祖先关系
+  }
+
+  /**
+   * 检查循环是否包含函数调用
+   * 使用场景: 内联决策、副作用分析
+   */
+  bool containsFunctionCalls() const {
+    for (BasicBlock* bb : LoopBlocks) {
+      for (auto& inst : bb->getInstructions()) {
+        if (dynamic_cast<CallInst*>(inst.get())) {
+          return true;
+        }
+      }
+    }
+    return false;
+  }
+
+  /**
+   * 检查循环是否可能有副作用（基于副作用分析结果）
+   * 使用场景: 循环优化决策、并行化分析
+   */
+  bool mayHaveSideEffects(SideEffectAnalysisResult* sideEffectAnalysis) const;
+
+  /**
+   * 检查循环是否访问全局内存（基于别名分析结果）
+   * 使用场景: 并行化分析、缓存优化
+   */
+  bool accessesGlobalMemory(AliasAnalysisResult* aliasAnalysis) const;
+
+  /**
+   * 检查循环是否有可能的内存别名冲突
+   * 使用场景: 向量化分析、并行化决策
+   */
+  bool hasMemoryAliasConflicts(AliasAnalysisResult* aliasAnalysis) const;
+
+  /**
+   * 估算循环的"热度" (基于嵌套深度和大小)
+   * 使用场景: 优化优先级、资源分配
+   */
+  double getLoopHotness() const {
+    // 简单的热度估算: 深度权重 + 大小惩罚
+    double hotness = std::pow(2.0, Level); // 深度越深越热
+    hotness /= std::sqrt(LoopBlocks.size()); // 大小越大相对热度降低
+    return hotness;
+  }
+
+  // --- 供 LoopAnalysisPass 内部调用的方法，用于构建 Loop 对象 ---
+  void addBlock(BasicBlock *BB) { LoopBlocks.insert(BB); }
+  void addExitBlock(BasicBlock *BB) { ExitBlocks.insert(BB); }
+  void setPreHeader(BasicBlock *BB) { PreHeader = BB; }
+  void setParentLoop(Loop *loop) { ParentLoop = loop; }
+  void addNestedLoop(Loop *loop) { NestedLoops.push_back(loop); }
+  void setLoopLevel(int level) { Level = level; }
+  void clearNestedLoops() { NestedLoops.clear(); }
+private:
+  BasicBlock *Header;                // 循环头基本块
+  std::set<BasicBlock *> LoopBlocks; // 循环体包含的基本块集合
+  std::set<BasicBlock *> ExitBlocks; // 循环出口基本块集合
+  BasicBlock *PreHeader = nullptr;   // 循环前置块 (Optional)
+  Loop *ParentLoop = nullptr;        // 父循环 (用于嵌套)
+  std::vector<Loop *> NestedLoops;   // 嵌套的子循环
+  int Level = -1;                    // 循环的层级，-1表示未计算
+};
+
+/**
+ * @brief 循环分析结果类。
+ * 包含一个函数中所有识别出的循环，并提供高效的查询缓存机制。
+ */
+class LoopAnalysisResult : public AnalysisResultBase {
+public:
+  LoopAnalysisResult(Function *F) : AssociatedFunction(F) {}
+  ~LoopAnalysisResult() override = default;
+
+  // ========== 缓存统计结构 ==========
+  struct CacheStats {
+    size_t innermostLoopsCached;
+    size_t outermostLoopsCached;
+    size_t loopsByDepthCached;
+    size_t containingLoopsCached;
+    size_t allNestedLoopsCached;
+    size_t totalCachedQueries;
+  };
+
+private:
+  // ========== 高频查询缓存 ==========
+  mutable std::optional<std::vector<Loop*>> cachedInnermostLoops;
+  mutable std::optional<std::vector<Loop*>> cachedOutermostLoops;
+  mutable std::optional<int> cachedMaxDepth;
+  mutable std::optional<size_t> cachedLoopCount;
+  mutable std::map<int, std::vector<Loop*>> cachedLoopsByDepth;
+  
+  // ========== 中频查询缓存 ==========
+  mutable std::map<BasicBlock*, Loop*> cachedInnermostContainingLoop;
+  mutable std::map<Loop*, std::set<Loop*>> cachedAllNestedLoops; // 递归嵌套
+  mutable std::map<BasicBlock*, std::vector<Loop*>> cachedAllContainingLoops;
+  
+  // ========== 缓存状态管理 ==========
+  mutable bool cacheValid = true;
+
+  // 内部辅助方法
+  void invalidateCache() const {
+    cachedInnermostLoops.reset();
+    cachedOutermostLoops.reset();
+    cachedMaxDepth.reset();
+    cachedLoopCount.reset();
+    cachedLoopsByDepth.clear();
+    cachedInnermostContainingLoop.clear();
+    cachedAllNestedLoops.clear();
+    cachedAllContainingLoops.clear();
+    cacheValid = false;
+  }
+  
+  void ensureCacheValid() const {
+    if (!cacheValid) {
+      // 重新计算基础缓存
+      computeBasicCache();
+      cacheValid = true;
+    }
+  }
+  
+  void computeBasicCache() const {
+    // 计算最内层循环
+    if (!cachedInnermostLoops) {
+      cachedInnermostLoops = std::vector<Loop*>();
+      for (const auto& loop : AllLoops) {
+        if (loop->isInnermost()) {
+          cachedInnermostLoops->push_back(loop.get());
+        }
+      }
+    }
+    
+    // 计算最外层循环
+    if (!cachedOutermostLoops) {
+      cachedOutermostLoops = std::vector<Loop*>();
+      for (const auto& loop : AllLoops) {
+        if (loop->isOutermost()) {
+          cachedOutermostLoops->push_back(loop.get());
+        }
+      }
+    }
+    
+    // 计算最大深度
+    if (!cachedMaxDepth) {
+      int maxDepth = 0;
+      for (const auto& loop : AllLoops) {
+        maxDepth = std::max(maxDepth, loop->getLoopDepth());
+      }
+      cachedMaxDepth = maxDepth;
+    }
+    
+    // 计算循环总数
+    if (!cachedLoopCount) {
+      cachedLoopCount = AllLoops.size();
+    }
+  }
+
+public:
+  // ========== 基础接口 ==========
+
+  // 添加一个识别出的循环到结果中
+  void addLoop(std::unique_ptr<Loop> loop) {
+    invalidateCache(); // 添加新循环时失效缓存
+    AllLoops.push_back(std::move(loop));
+    LoopMap[AllLoops.back()->getHeader()] = AllLoops.back().get();
+  }
+
+  // 获取所有识别出的循环（unique_ptr 管理内存）
+  const std::vector<std::unique_ptr<Loop>> &getAllLoops() const { return AllLoops; }
+
+  // ========== 高频查询接口 ==========
+  
+  /**
+   * 获取所有最内层循环 - 循环优化的主要目标
+   * 使用场景: 循环展开、向量化、循环不变量外提
+   */
+  const std::vector<Loop*>& getInnermostLoops() const {
+    ensureCacheValid();
+    if (!cachedInnermostLoops) {
+      cachedInnermostLoops = std::vector<Loop*>();
+      for (const auto& loop : AllLoops) {
+        if (loop->isInnermost()) {
+          cachedInnermostLoops->push_back(loop.get());
+        }
+      }
+    }
+    return *cachedInnermostLoops;
+  }
+  
+  /**
+   * 获取所有最外层循环
+   * 使用场景: 循环树遍历、整体优化策略
+   */
+  const std::vector<Loop*>& getOutermostLoops() const {
+    ensureCacheValid();
+    if (!cachedOutermostLoops) {
+      cachedOutermostLoops = std::vector<Loop*>();
+      for (const auto& loop : AllLoops) {
+        if (loop->isOutermost()) {
+          cachedOutermostLoops->push_back(loop.get());
+        }
+      }
+    }
+    return *cachedOutermostLoops;
+  }
+  
+  /**
+   * 获取指定深度的所有循环
+   * 使用场景: 分层优化、循环展开决策、并行化分析
+   */
+  const std::vector<Loop*>& getLoopsAtDepth(int depth) const {
+    ensureCacheValid();
+    if (cachedLoopsByDepth.find(depth) == cachedLoopsByDepth.end()) {
+      std::vector<Loop*> result;
+      for (const auto& loop : AllLoops) {
+        if (loop->getLoopDepth() == depth) {
+          result.push_back(loop.get());
+        }
+      }
+      cachedLoopsByDepth[depth] = std::move(result);
+    }
+    return cachedLoopsByDepth[depth];
+  }
+  
+  /**
+   * 获取最大循环嵌套深度
+   * 使用场景: 优化预算分配、编译时间控制
+   */
+  int getMaxLoopDepth() const {
+    ensureCacheValid();
+    if (!cachedMaxDepth) {
+      int maxDepth = 0;
+      for (const auto& loop : AllLoops) {
+        maxDepth = std::max(maxDepth, loop->getLoopDepth());
+      }
+      cachedMaxDepth = maxDepth;
+    }
+    return *cachedMaxDepth;
+  }
+  
+  /**
+   * 获取循环总数
+   * 使用场景: 统计信息、优化决策
+   */
+  size_t getLoopCount() const {
+    ensureCacheValid();
+    if (!cachedLoopCount) {
+      cachedLoopCount = AllLoops.size();
+    }
+    return *cachedLoopCount;
+  }
+
+  // 获取指定深度的循环数量
+  size_t getLoopCountAtDepth(int depth) const {
+    return getLoopsAtDepth(depth).size();
+  }
+
+  // 检查函数是否包含循环
+  bool hasLoops() const { return !AllLoops.empty(); }
+
+  // ========== 中频查询接口 ==========
+  
+  /**
+   * 获取包含指定基本块的最内层循环
+   * 使用场景: 活跃性分析、寄存器分配、指令调度
+   */
+  Loop* getInnermostContainingLoop(BasicBlock* BB) const {
+    ensureCacheValid();
+    if (cachedInnermostContainingLoop.find(BB) == cachedInnermostContainingLoop.end()) {
+      Loop* result = nullptr;
+      int maxDepth = -1;
+      for (const auto& loop : AllLoops) {
+        if (loop->contains(BB) && loop->getLoopDepth() > maxDepth) {
+          result = loop.get();
+          maxDepth = loop->getLoopDepth();
+        }
+      }
+      cachedInnermostContainingLoop[BB] = result;
+    }
+    return cachedInnermostContainingLoop[BB];
+  }
+  
+  /**
+   * 获取包含指定基本块的所有循环 (从外到内排序)
+   * 使用场景: 循环间优化、依赖分析
+   */
+  const std::vector<Loop*>& getAllContainingLoops(BasicBlock* BB) const {
+    ensureCacheValid();
+    if (cachedAllContainingLoops.find(BB) == cachedAllContainingLoops.end()) {
+      std::vector<Loop*> result;
+      for (const auto& loop : AllLoops) {
+        if (loop->contains(BB)) {
+          result.push_back(loop.get());
+        }
+      }
+      // 按深度排序 (外层到内层)
+      std::sort(result.begin(), result.end(), 
+                [](Loop* a, Loop* b) { return a->getLoopDepth() < b->getLoopDepth(); });
+      cachedAllContainingLoops[BB] = std::move(result);
+    }
+    return cachedAllContainingLoops[BB];
+  }
+  
+  /**
+   * 获取指定循环的所有嵌套子循环 (递归)
+   * 使用场景: 循环树分析、嵌套优化
+   */
+  const std::set<Loop*>& getAllNestedLoops(Loop* loop) const {
+    ensureCacheValid();
+    if (cachedAllNestedLoops.find(loop) == cachedAllNestedLoops.end()) {
+      std::set<Loop*> result;
+      std::function<void(Loop*)> collectNested = [&](Loop* current) {
+        for (Loop* nested : current->getNestedLoops()) {
+          result.insert(nested);
+          collectNested(nested); // 递归收集
+        }
+      };
+      collectNested(loop);
+      cachedAllNestedLoops[loop] = std::move(result);
+    }
+    return cachedAllNestedLoops[loop];
+  }
+
+  // ========== 利用别名和副作用分析的查询接口 ==========
+  
+  /**
+   * 获取所有纯循环（无副作用的循环）
+   * 并行化、循环优化
+   */
+  std::vector<Loop*> getPureLoops(SideEffectAnalysisResult* sideEffectAnalysis) const {
+    std::vector<Loop*> result;
+    if (!sideEffectAnalysis) return result;
+    
+    for (const auto& loop : AllLoops) {
+      if (!loop->mayHaveSideEffects(sideEffectAnalysis)) {
+        result.push_back(loop.get());
+      }
+    }
+    return result;
+  }
+  
+  /**
+   * 获取所有只访问局部内存的循环
+   * 缓存优化、局部性分析
+   */
+  std::vector<Loop*> getLocalMemoryLoops(AliasAnalysisResult* aliasAnalysis) const {
+    std::vector<Loop*> result;
+    if (!aliasAnalysis) return result;
+    
+    for (const auto& loop : AllLoops) {
+      if (!loop->accessesGlobalMemory(aliasAnalysis)) {
+        result.push_back(loop.get());
+      }
+    }
+    return result;
+  }
+  
+  /**
+   * 获取所有无内存别名冲突的循环
+   * 向量化、并行化
+   */
+  std::vector<Loop*> getNoAliasConflictLoops(AliasAnalysisResult* aliasAnalysis) const {
+    std::vector<Loop*> result;
+    if (!aliasAnalysis) return result;
+    
+    for (const auto& loop : AllLoops) {
+      if (!loop->hasMemoryAliasConflicts(aliasAnalysis)) {
+        result.push_back(loop.get());
+      }
+    }
+    return result;
+  }
+
+  // ========== 低频查询接口(不缓存) ==========
+  
+  /**
+   * 检查两个循环是否有嵌套关系
+   * 循环间依赖分析
+   */
+  bool isNestedLoop(Loop* inner, Loop* outer) const {
+    if (inner == outer) return false;
+    
+    Loop* current = inner->getParentLoop();
+    while (current) {
+      if (current == outer) return true;
+      current = current->getParentLoop();
+    }
+    return false;
+  }
+  
+  /**
+   * 获取两个循环的最近公共祖先循环
+   * 循环融合分析、优化范围确定
+   */
+  Loop* getLowestCommonAncestor(Loop* loop1, Loop* loop2) const {
+    if (!loop1 || !loop2) return nullptr;
+    if (loop1 == loop2) return loop1;
+    
+    // 收集loop1的所有祖先
+    std::set<Loop*> ancestors1;
+    Loop* current = loop1;
+    while (current) {
+      ancestors1.insert(current);
+      current = current->getParentLoop();
+    }
+    
+    // 查找loop2祖先链中第一个在ancestors1中的循环
+    current = loop2;
+    while (current) {
+      if (ancestors1.count(current)) {
+        return current;
+      }
+      current = current->getParentLoop();
+    }
+    
+    return nullptr; // 没有公共祖先
+  }
+
+  // 通过循环头获取 Loop 对象
+  Loop *getLoopForHeader(BasicBlock *header) const {
+    auto it = LoopMap.find(header);
+    return (it != LoopMap.end()) ? it->second : nullptr;
+  }
+
+  // 通过某个基本块获取包含它的最内层循环 (向后兼容接口)
+  Loop *getLoopContainingBlock(BasicBlock *BB) const {
+    return getInnermostContainingLoop(BB);
+  }
+
+  // ========== 缓存管理接口 ==========
+  
+  /**
+   * 手动失效缓存 (可删除)
+   */
+  void invalidateQueryCache() const {
+    invalidateCache();
+  }
+  
+  /**
+   * 获取缓存统计信息
+   */
+  CacheStats getCacheStats() const {
+    CacheStats stats = {};
+    stats.innermostLoopsCached = cachedInnermostLoops.has_value() ? 1 : 0;
+    stats.outermostLoopsCached = cachedOutermostLoops.has_value() ? 1 : 0;
+    stats.loopsByDepthCached = cachedLoopsByDepth.size();
+    stats.containingLoopsCached = cachedInnermostContainingLoop.size();
+    stats.allNestedLoopsCached = cachedAllNestedLoops.size();
+    stats.totalCachedQueries = stats.innermostLoopsCached + stats.outermostLoopsCached + 
+                               stats.loopsByDepthCached + stats.containingLoopsCached + 
+                               stats.allNestedLoopsCached;
+    return stats;
+  }
+
+  // 打印分析结果
+  void print() const;
+  void printBBSet(const std::string &prefix, const std::set<BasicBlock *> &s) const;
+  void printLoopVector(const std::string &prefix, const std::vector<Loop *> &loops) const;
+
+private:
+  Function *AssociatedFunction;                // 结果关联的函数
+  std::vector<std::unique_ptr<Loop>> AllLoops; // 所有识别出的循环
+  std::map<BasicBlock *, Loop *> LoopMap;      // 循环头到 Loop* 的映射，方便查找
+};
+
+/**
+ * @brief 循环分析遍。
+ * 识别函数中的所有循环，并生成 LoopAnalysisResult。
+ */
+class LoopAnalysisPass : public AnalysisPass {
+public:
+  // 唯一的 Pass ID，需要在 .cpp 文件中定义
+  static void *ID;
+
+  LoopAnalysisPass() : AnalysisPass("LoopAnalysis", Pass::Granularity::Function) {}
+
+  // 实现 getPassID
+  void *getPassID() const override { return &ID; }
+
+  // 核心运行方法：在每个函数上执行循环分析
+  bool runOnFunction(Function *F, AnalysisManager &AM) override;
+
+  // 获取分析结果
+  std::unique_ptr<AnalysisResultBase> getResult() override { return std::move(CurrentResult); }
+
+private:
+  std::unique_ptr<LoopAnalysisResult> CurrentResult; // 当前函数的分析结果
+};
+
+} // namespace sysy

src/include/midend/Pass/Analysis/LoopCharacteristics.h

@@ -0,0 +1,360 @@
+#pragma once
+
+#include "Dom.h"         // 支配树分析依赖
+#include "Loop.h"        // 循环分析依赖
+#include "Liveness.h"    // 活跃性分析依赖
+#include "AliasAnalysis.h" // 别名分析依赖
+#include "SideEffectAnalysis.h" // 副作用分析依赖
+#include "CallGraphAnalysis.h" // 调用图分析依赖
+#include "IR.h"          // IR定义
+#include "Pass.h"        // Pass框架
+#include <algorithm>
+#include <map>
+#include <memory>
+#include <optional>
+#include <set>
+#include <vector>
+
+namespace sysy {
+
+// 前向声明
+class LoopCharacteristicsResult;
+
+enum IVKind {
+  kBasic,        // 基本归纳变量
+  kLinear,     // 线性归纳变量
+  kCmplx       // 复杂派生归纳变量
+} ;          // 归纳变量类型
+
+struct InductionVarInfo {
+  Value* div;                  // 派生归纳变量的指令
+  Value* base = nullptr;                 // 其根phi或BIV或DIV
+  std::pair<Value*, Value*> Multibase = {nullptr, nullptr}; // 多个BIV
+  Instruction::Kind Instkind;      // 操作类型
+  int factor = 1;              // 系数（如i*2+3的2）
+  int offset = 0;              // 常量偏移
+  bool valid;    // 是否线性可归约
+  IVKind ivkind;          // 归纳变量类型
+
+
+static std::unique_ptr<InductionVarInfo> createBasicBIV(Value* v, Instruction::Kind kind, Value* base = nullptr, int factor = 1, int offset = 0) {
+  return std::make_unique<InductionVarInfo>(
+    InductionVarInfo{v, base, {nullptr, nullptr}, kind, factor, offset, true, IVKind::kBasic}
+  );
+}
+
+static std::unique_ptr<InductionVarInfo> createSingleDIV(Value* v, Instruction::Kind kind, Value* base = nullptr, int factor = 1, int offset = 0) {
+  return std::make_unique<InductionVarInfo>(
+    InductionVarInfo{v, base, {nullptr, nullptr}, kind, factor, offset, true,  IVKind::kLinear}
+  );
+}
+
+static std::unique_ptr<InductionVarInfo> createDoubleDIV(Value* v, Instruction::Kind kind, Value* base1 = nullptr,  Value* base2 = nullptr, int factor = 1, int offset = 0) {
+  return std::make_unique<InductionVarInfo>(
+    InductionVarInfo{v, nullptr, {base1, base2}, kind, factor, offset, false, IVKind::kCmplx}
+  );
+}
+};
+
+/**
+ * @brief 循环特征信息结构 - 基础循环分析阶段
+ * 存储循环的基本特征信息，为后续精确分析提供基础
+ */
+struct LoopCharacteristics {
+  Loop* loop;                                    // 关联的循环对象
+
+  // ========== 基础循环形式分析 ==========
+  bool isCountingLoop;                         // 是否为计数循环 (for i=0; i<n; i++)
+  bool isSimpleForLoop;                        // 是否为简单for循环
+  bool hasComplexControlFlow;                  // 是否有复杂控制流 (break, continue)
+  bool isInnermost;                            // 是否为最内层循环
+  
+  // ========== 归纳变量分析 ==========
+  
+  // ========== 基础循环不变量分析 ==========
+  std::unordered_set<Value*> loopInvariants;              // 循环不变量
+  std::unordered_set<Instruction*> invariantInsts;       // 可提升的不变指令
+
+  std::vector<std::unique_ptr<InductionVarInfo>> InductionVars;     // 归纳变量
+  
+  // ========== 基础边界分析 ==========
+  std::optional<int> staticTripCount;          // 静态循环次数（如果可确定）
+  bool hasKnownBounds;                         // 是否有已知边界
+  
+  // ========== 基础纯度和副作用分析 ==========
+  bool isPure;                                 // 是否为纯循环（无副作用）
+  bool accessesOnlyLocalMemory;                // 是否只访问局部内存
+  bool hasNoMemoryAliasConflicts;              // 是否无内存别名冲突
+  
+  // ========== 基础内存访问模式分析 ==========
+  struct MemoryAccessPattern {
+    std::vector<Instruction*> loadInsts;       // load指令列表
+    std::vector<Instruction*> storeInsts;      // store指令列表
+    bool isArrayParameter;                     // 是否为数组参数访问
+    bool isGlobalArray;                        // 是否为全局数组访问
+    bool hasConstantIndices;                   // 是否使用常量索引
+  };
+  std::map<Value*, MemoryAccessPattern> memoryPatterns; // 内存访问模式
+  
+  // ========== 基础性能特征 ==========
+  size_t instructionCount;                     // 循环体指令数
+  size_t memoryOperationCount;                 // 内存操作数
+  size_t arithmeticOperationCount;             // 算术操作数
+  double computeToMemoryRatio;                 // 计算与内存操作比率
+  
+  // ========== 基础优化提示 ==========
+  bool benefitsFromUnrolling;                  // 是否适合循环展开
+  int suggestedUnrollFactor;                   // 建议的展开因子
+  
+  // 构造函数 - 简化的基础分析初始化
+  LoopCharacteristics(Loop* l) : loop(l), 
+    isCountingLoop(false), isSimpleForLoop(false), hasComplexControlFlow(false),
+    isInnermost(false), hasKnownBounds(false), isPure(false), 
+    accessesOnlyLocalMemory(false), hasNoMemoryAliasConflicts(false),
+    benefitsFromUnrolling(false), suggestedUnrollFactor(1), 
+    instructionCount(0), memoryOperationCount(0),
+    arithmeticOperationCount(0), computeToMemoryRatio(0.0) {}
+};
+
+/**
+ * @brief 循环特征分析结果类
+ * 包含函数中所有循环的特征信息，并提供查询接口
+ */
+class LoopCharacteristicsResult : public AnalysisResultBase {
+public:
+  LoopCharacteristicsResult(Function *F) : AssociatedFunction(F) {}
+  ~LoopCharacteristicsResult() override = default;
+
+  // ========== 基础接口 ==========
+  
+  /**
+   * 添加循环特征信息
+   */
+  void addLoopCharacteristics(std::unique_ptr<LoopCharacteristics> characteristics) {
+    auto* loop = characteristics->loop;
+    CharacteristicsMap[loop] = std::move(characteristics);
+  }
+  
+  /**
+   * 获取指定循环的特征信息
+   */
+  const LoopCharacteristics* getCharacteristics(Loop* loop) const {
+    auto it = CharacteristicsMap.find(loop);
+    return (it != CharacteristicsMap.end()) ? it->second.get() : nullptr;
+  }
+  
+  /**
+   * 获取所有循环特征信息
+   */
+  const std::map<Loop*, std::unique_ptr<LoopCharacteristics>>& getAllCharacteristics() const {
+    return CharacteristicsMap;
+  }
+
+  // ========== 核心查询接口 ==========
+  
+  /**
+   * 获取所有计数循环
+   */
+  std::vector<Loop*> getCountingLoops() const {
+    std::vector<Loop*> result;
+    for (const auto& [loop, chars] : CharacteristicsMap) {
+      if (chars->isCountingLoop) {
+        result.push_back(loop);
+      }
+    }
+    return result;
+  }
+  
+  /**
+   * 获取所有纯循环（无副作用）
+   */
+  std::vector<Loop*> getPureLoops() const {
+    std::vector<Loop*> result;
+    for (const auto& [loop, chars] : CharacteristicsMap) {
+      if (chars->isPure) {
+        result.push_back(loop);
+      }
+    }
+    return result;
+  }
+  
+  /**
+   * 获取所有只访问局部内存的循环
+   */
+  std::vector<Loop*> getLocalMemoryOnlyLoops() const {
+    std::vector<Loop*> result;
+    for (const auto& [loop, chars] : CharacteristicsMap) {
+      if (chars->accessesOnlyLocalMemory) {
+        result.push_back(loop);
+      }
+    }
+    return result;
+  }
+  
+  /**
+   * 获取所有无内存别名冲突的循环
+   */
+  std::vector<Loop*> getNoAliasConflictLoops() const {
+    std::vector<Loop*> result;
+    for (const auto& [loop, chars] : CharacteristicsMap) {
+      if (chars->hasNoMemoryAliasConflicts) {
+        result.push_back(loop);
+      }
+    }
+    return result;
+  }
+  
+  /**
+   * 获取所有适合展开的循环
+   */
+  std::vector<Loop*> getUnrollingCandidates() const {
+    std::vector<Loop*> result;
+    for (const auto& [loop, chars] : CharacteristicsMap) {
+      if (chars->benefitsFromUnrolling) {
+        result.push_back(loop);
+      }
+    }
+    return result;
+  }
+  
+  /**
+   * 根据热度排序循环 (用于优化优先级)
+   */
+  std::vector<Loop*> getLoopsByHotness() const {
+    std::vector<Loop*> result;
+    for (const auto& [loop, chars] : CharacteristicsMap) {
+      result.push_back(loop);
+    }
+    
+    // 按循环热度排序 (嵌套深度 + 循环次数 + 指令数)
+    std::sort(result.begin(), result.end(), [](Loop* a, Loop* b) {
+      double hotnessA = a->getLoopHotness();
+      double hotnessB = b->getLoopHotness();
+      return hotnessA > hotnessB; // 降序排列
+    });
+    
+    return result;
+  }
+
+  // ========== 基础统计接口 ==========
+  
+  /**
+   * 获取基础优化统计信息
+   */
+  struct BasicOptimizationStats {
+    size_t totalLoops;
+    size_t countingLoops;
+    size_t unrollingCandidates;
+    size_t pureLoops;
+    size_t localMemoryOnlyLoops;
+    size_t noAliasConflictLoops;
+    double avgInstructionCount;
+    double avgComputeMemoryRatio;
+  };
+  
+  BasicOptimizationStats getOptimizationStats() const {
+    BasicOptimizationStats stats = {};
+    stats.totalLoops = CharacteristicsMap.size();
+    
+    size_t totalInstructions = 0;
+    double totalComputeMemoryRatio = 0.0;
+    
+    for (const auto& [loop, chars] : CharacteristicsMap) {
+      if (chars->isCountingLoop) stats.countingLoops++;
+      if (chars->benefitsFromUnrolling) stats.unrollingCandidates++;
+      if (chars->isPure) stats.pureLoops++;
+      if (chars->accessesOnlyLocalMemory) stats.localMemoryOnlyLoops++;
+      if (chars->hasNoMemoryAliasConflicts) stats.noAliasConflictLoops++;
+      
+      totalInstructions += chars->instructionCount;
+      totalComputeMemoryRatio += chars->computeToMemoryRatio;
+    }
+    
+    if (stats.totalLoops > 0) {
+      stats.avgInstructionCount = static_cast<double>(totalInstructions) / stats.totalLoops;
+      stats.avgComputeMemoryRatio = totalComputeMemoryRatio / stats.totalLoops;
+    }
+    
+    return stats;
+  }
+
+  // 打印分析结果
+  void print() const;
+
+private:
+  Function *AssociatedFunction;                                    // 关联的函数
+  std::map<Loop*, std::unique_ptr<LoopCharacteristics>> CharacteristicsMap; // 循环特征映射
+};
+
+/**
+ * @brief 基础循环特征分析遍
+ * 在循环规范化前执行，进行基础的循环特征分析，为后续精确分析提供基础
+ */
+class LoopCharacteristicsPass : public AnalysisPass {
+public:
+  // 唯一的 Pass ID
+  static void *ID;
+
+  LoopCharacteristicsPass() : AnalysisPass("LoopCharacteristics", Pass::Granularity::Function) {}
+
+  // 实现 getPassID
+  void *getPassID() const override { return &ID; }
+
+  // 核心运行方法
+  bool runOnFunction(Function *F, AnalysisManager &AM) override;
+
+  // 获取分析结果
+  std::unique_ptr<AnalysisResultBase> getResult() override { return std::move(CurrentResult); }
+
+private:
+  std::unique_ptr<LoopCharacteristicsResult> CurrentResult;
+  
+  // ========== 缓存的分析结果 ==========
+  LoopAnalysisResult* loopAnalysis;           // 循环结构分析结果
+  AliasAnalysisResult* aliasAnalysis;         // 别名分析结果  
+  SideEffectAnalysisResult* sideEffectAnalysis; // 副作用分析结果
+  
+  // ========== 核心分析方法 ==========
+  void analyzeLoop(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 基础循环形式分析
+  void analyzeLoopForm(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 基础性能指标计算
+  void computePerformanceMetrics(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 基础纯度和副作用分析
+  void analyzePurityAndSideEffects(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 基础归纳变量识别
+  void identifyBasicInductionVariables(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 循环不变量识别
+  void identifyBasicLoopInvariants(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 基础边界分析
+  void analyzeBasicLoopBounds(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 基础内存访问模式分析
+  void analyzeBasicMemoryAccessPatterns(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // 基础优化评估
+  void evaluateBasicOptimizationOpportunities(Loop* loop, LoopCharacteristics* characteristics);
+  
+  // ========== 辅助方法 ==========
+  bool isClassicLoopInvariant(Value* val, Loop* loop, const std::unordered_set<Value*>& invariants);
+  void findDerivedInductionVars(Value* root,
+    Value* base, // 只传单一BIV base
+    Loop* loop,
+    std::vector<std::unique_ptr<InductionVarInfo>>& ivs,
+    std::set<Value*>& visited
+  );
+  bool isBasicInductionVariable(Value* val, Loop* loop);
+  // ========== 循环不变量分析辅助方法 ==========
+  bool isInvariantOperands(Instruction* inst, Loop* loop, const std::unordered_set<Value*>& invariants);
+  bool isMemoryLocationModifiedInLoop(Value* ptr, Loop* loop);
+  bool isMemoryLocationLoadedInLoop(Value* ptr, Loop* loop, Instruction* excludeInst = nullptr);
+  bool isPureFunction(Function* calledFunc);
+};
+
+} // namespace sysy

src/include/midend/Pass/Analysis/LoopVectorization.h

@@ -0,0 +1,250 @@
+#pragma once
+
+#include "Pass.h"
+#include "Loop.h" 
+#include "LoopCharacteristics.h"
+#include "AliasAnalysis.h"
+#include "SideEffectAnalysis.h"
+#include <vector>
+#include <map>
+#include <memory>
+#include <set>
+#include <string>
+
+namespace sysy {
+
+/**
+ * @brief 依赖类型枚举 - 只考虑真正影响并行性的依赖
+ * 
+ * 依赖类型分析说明：
+ * - TRUE_DEPENDENCE (RAW): 真依赖，必须保持原始执行顺序，是最关键的依赖
+ * - ANTI_DEPENDENCE (WAR): 反依赖，影响指令重排序，可通过寄存器重命名等技术缓解
+ * - OUTPUT_DEPENDENCE (WAW): 输出依赖，相对较少但需要考虑，可通过变量私有化解决
+ * 
+ */
+enum class DependenceType {
+  TRUE_DEPENDENCE,    // 真依赖 (RAW) - 读后写流依赖，最重要的依赖类型
+  ANTI_DEPENDENCE,    // 反依赖 (WAR) - 写后读反向依赖，影响指令重排序  
+  OUTPUT_DEPENDENCE   // 输出依赖 (WAW) - 写后写，相对较少但需要考虑
+};
+
+/**
+ * @brief 依赖向量 - 表示两个内存访问之间的迭代距离
+ * 例如：a[i] 和 a[i+1] 之间的依赖向量是 [1]
+ *       a[i][j] 和 a[i+1][j-2] 之间的依赖向量是 [1,-2]
+ */
+struct DependenceVector {
+  std::vector<int> distances;    // 每个循环层次的依赖距离
+  bool isConstant;               // 是否为常量距离
+  bool isKnown;                  // 是否已知距离
+  
+  DependenceVector(size_t loopDepth) : distances(loopDepth, 0), isConstant(false), isKnown(false) {}
+  
+  // 检查是否为循环无关依赖
+  bool isLoopIndependent() const {
+    for (int dist : distances) {
+      if (dist != 0) return false;
+    }
+    return true;
+  }
+  
+  // 获取词典序方向向量
+  std::vector<int> getDirectionVector() const;
+  
+  // 检查是否可以通过向量化处理
+  bool isVectorizationSafe() const;
+};
+
+/**
+ * @brief 精确依赖关系 - 包含依赖向量的详细依赖信息
+ */
+struct PreciseDependence {
+  Instruction* source;
+  Instruction* sink;
+  DependenceType type;
+  DependenceVector dependenceVector;
+  Value* memoryLocation;
+  
+  // 并行化相关
+  bool allowsParallelization;    // 是否允许并行化
+  bool requiresSynchronization;  // 是否需要同步
+  bool isReductionDependence;    // 是否为归约依赖
+  
+  PreciseDependence(size_t loopDepth) : dependenceVector(loopDepth), 
+    allowsParallelization(true), requiresSynchronization(false), isReductionDependence(false) {}
+};
+
+/**
+ * @brief 向量化分析信息 - 暂时搁置，保留接口
+ */
+struct VectorizationAnalysis {
+  bool isVectorizable;                        // 固定为false，暂不支持
+  int suggestedVectorWidth;                   // 固定为1
+  std::vector<std::string> preventingFactors; // 阻止向量化的因素
+  
+  VectorizationAnalysis() : isVectorizable(false), suggestedVectorWidth(1) {
+    preventingFactors.push_back("Vectorization temporarily disabled");
+  }
+};
+
+/**
+ * @brief 并行化分析信息
+ */
+struct ParallelizationAnalysis {
+  bool isParallelizable;                     // 是否可并行化
+  int suggestedThreadCount;                  // 建议的线程数
+  std::vector<std::string> preventingFactors; // 阻止并行化的因素
+  
+  // 并行化模式
+  enum ParallelizationType { 
+    NONE,                    // 不可并行化
+    EMBARRASSINGLY_PARALLEL, // 完全并行
+    REDUCTION_PARALLEL,      // 归约并行
+    PIPELINE_PARALLEL,       // 流水线并行
+    CONDITIONAL_PARALLEL     // 条件并行
+  } parallelType;
+  
+  // 负载均衡
+  bool hasLoadBalance;                       // 是否有良好的负载均衡
+  bool isDynamicLoadBalanced;                // 是否需要动态负载均衡
+  double workComplexity;                     // 工作复杂度估计
+  
+  // 同步需求
+  bool requiresReduction;                    // 是否需要归约操作
+  bool requiresBarrier;                      // 是否需要屏障同步
+  std::set<Value*> sharedVariables;         // 共享变量
+  std::set<Value*> reductionVariables;      // 归约变量
+  std::set<Value*> privatizableVariables;   // 可私有化变量
+  
+  // 内存访问模式
+  bool hasMemoryConflicts;                   // 是否有内存冲突
+  bool hasReadOnlyAccess;                    // 是否只有只读访问
+  bool hasIndependentAccess;                 // 是否有独立的内存访问
+  
+  // 并行化收益评估
+  double parallelizationBenefit;             // 并行化收益估计 (0-1)
+  size_t communicationCost;                  // 通信开销估计
+  size_t synchronizationCost;                // 同步开销估计
+  
+  ParallelizationAnalysis() : isParallelizable(false), suggestedThreadCount(1), parallelType(NONE),
+    hasLoadBalance(true), isDynamicLoadBalanced(false), workComplexity(0.0), requiresReduction(false), 
+    requiresBarrier(false), hasMemoryConflicts(false), hasReadOnlyAccess(false), hasIndependentAccess(false),
+    parallelizationBenefit(0.0), communicationCost(0), synchronizationCost(0) {}
+};
+
+/**
+ * @brief 循环向量化/并行化分析结果
+ */
+class LoopVectorizationResult : public AnalysisResultBase {
+private:
+  Function* AssociatedFunction;
+  std::map<Loop*, VectorizationAnalysis> VectorizationMap;
+  std::map<Loop*, ParallelizationAnalysis> ParallelizationMap;
+  std::map<Loop*, std::vector<PreciseDependence>> DependenceMap;
+
+public:
+  LoopVectorizationResult(Function* F) : AssociatedFunction(F) {}
+  ~LoopVectorizationResult() override = default;
+
+  // 基础接口
+  void addVectorizationAnalysis(Loop* loop, VectorizationAnalysis analysis) {
+    VectorizationMap[loop] = std::move(analysis);
+  }
+  
+  void addParallelizationAnalysis(Loop* loop, ParallelizationAnalysis analysis) {
+    ParallelizationMap[loop] = std::move(analysis);
+  }
+  
+  void addDependenceAnalysis(Loop* loop, std::vector<PreciseDependence> dependences) {
+    DependenceMap[loop] = std::move(dependences);
+  }
+
+  // 查询接口
+  const VectorizationAnalysis* getVectorizationAnalysis(Loop* loop) const {
+    auto it = VectorizationMap.find(loop);
+    return it != VectorizationMap.end() ? &it->second : nullptr;
+  }
+  
+  const ParallelizationAnalysis* getParallelizationAnalysis(Loop* loop) const {
+    auto it = ParallelizationMap.find(loop);
+    return it != ParallelizationMap.end() ? &it->second : nullptr;
+  }
+  
+  const std::vector<PreciseDependence>* getPreciseDependences(Loop* loop) const {
+    auto it = DependenceMap.find(loop);
+    return it != DependenceMap.end() ? &it->second : nullptr;
+  }
+
+  // 统计接口
+  size_t getVectorizableLoopCount() const;
+  size_t getParallelizableLoopCount() const;
+  
+  // 优化建议
+  std::vector<Loop*> getVectorizationCandidates() const;
+  std::vector<Loop*> getParallelizationCandidates() const;
+  
+  // 打印分析结果
+  void print() const;
+};
+
+/**
+ * @brief 循环向量化/并行化分析遍
+ * 在循环规范化后执行，进行精确的依赖向量分析和向量化/并行化可行性评估
+ * 专注于并行化分析，向量化功能暂时搁置
+ */
+class LoopVectorizationPass : public AnalysisPass {
+public:
+  // 唯一的 Pass ID
+  static void *ID;
+  
+  LoopVectorizationPass() : AnalysisPass("LoopVectorization", Pass::Granularity::Function) {}
+
+  // 实现 getPassID
+  void *getPassID() const override { return &ID; }
+
+  // 核心运行方法
+  bool runOnFunction(Function *F, AnalysisManager &AM) override;
+  
+  // 获取分析结果
+  std::unique_ptr<AnalysisResultBase> getResult() override { return std::move(CurrentResult); }
+
+private:
+  std::unique_ptr<LoopVectorizationResult> CurrentResult;
+  
+  // ========== 主要分析方法 ==========
+  void analyzeLoop(Loop* loop, LoopCharacteristics* characteristics, 
+                   AliasAnalysisResult* aliasAnalysis, SideEffectAnalysisResult* sideEffectAnalysis);
+  
+  // ========== 依赖向量分析 ==========
+  std::vector<PreciseDependence> computeDependenceVectors(Loop* loop, AliasAnalysisResult* aliasAnalysis);
+  DependenceVector computeAccessDependence(Instruction* inst1, Instruction* inst2, Loop* loop);
+  bool areAccessesAffinelyRelated(Value* ptr1, Value* ptr2, Loop* loop);
+  
+  // ========== 向量化分析 (暂时搁置) ==========
+  VectorizationAnalysis analyzeVectorizability(Loop* loop, const std::vector<PreciseDependence>& dependences,
+                                              LoopCharacteristics* characteristics);
+  
+  // ========== 并行化分析 ==========
+  ParallelizationAnalysis analyzeParallelizability(Loop* loop, const std::vector<PreciseDependence>& dependences,
+                                                  LoopCharacteristics* characteristics);
+  bool checkParallelizationLegality(Loop* loop, const std::vector<PreciseDependence>& dependences);
+  int estimateOptimalThreadCount(Loop* loop, LoopCharacteristics* characteristics);
+  ParallelizationAnalysis::ParallelizationType determineParallelizationType(Loop* loop, 
+                                                                           const std::vector<PreciseDependence>& dependences);
+  
+  // ========== 并行化专用分析方法 ==========
+  void analyzeReductionPatterns(Loop* loop, ParallelizationAnalysis* analysis);
+  void analyzeMemoryAccessPatterns(Loop* loop, ParallelizationAnalysis* analysis, AliasAnalysisResult* aliasAnalysis);
+  void estimateParallelizationBenefit(Loop* loop, ParallelizationAnalysis* analysis, LoopCharacteristics* characteristics);
+  void identifyPrivatizableVariables(Loop* loop, ParallelizationAnalysis* analysis);
+  void analyzeSynchronizationNeeds(Loop* loop, ParallelizationAnalysis* analysis, const std::vector<PreciseDependence>& dependences);
+  
+  // ========== 辅助方法 ==========
+  std::vector<int> extractInductionCoefficients(Value* ptr, Loop* loop);
+  bool isConstantStride(Value* ptr, Loop* loop, int& stride);
+  bool isIndependentMemoryAccess(Value* ptr1, Value* ptr2, Loop* loop);
+  double estimateWorkComplexity(Loop* loop);
+  bool hasReductionPattern(Value* var, Loop* loop);
+};
+
+} // namespace sysy

src/include/midend/Pass/Analysis/SideEffectAnalysis.h

@@ -0,0 +1,137 @@
+#pragma once
+
+#include "Pass.h"
+#include "IR.h"
+#include "AliasAnalysis.h"
+#include "CallGraphAnalysis.h"
+#include <unordered_set>
+#include <unordered_map>
+
+namespace sysy {
+
+// 副作用类型枚举
+enum class SideEffectType {
+    NO_SIDE_EFFECT,      // 无副作用
+    MEMORY_WRITE,        // 内存写入（store、memset）
+    FUNCTION_CALL,       // 函数调用（可能有任意副作用）
+    IO_OPERATION,        // I/O操作（printf、scanf等）
+    UNKNOWN              // 未知副作用
+};
+
+// 副作用信息结构
+struct SideEffectInfo {
+    SideEffectType type = SideEffectType::NO_SIDE_EFFECT;
+    bool mayModifyGlobal = false;      // 可能修改全局变量
+    bool mayModifyMemory = false;      // 可能修改内存
+    bool mayCallFunction = false;      // 可能调用函数
+    bool isPure = true;                // 是否为纯函数（无副作用且结果只依赖参数）
+    
+    // 合并两个副作用信息
+    SideEffectInfo merge(const SideEffectInfo& other) const {
+        SideEffectInfo result;
+        result.type = (type == SideEffectType::NO_SIDE_EFFECT) ? other.type : type;
+        result.mayModifyGlobal = mayModifyGlobal || other.mayModifyGlobal;
+        result.mayModifyMemory = mayModifyMemory || other.mayModifyMemory;
+        result.mayCallFunction = mayCallFunction || other.mayCallFunction;
+        result.isPure = isPure && other.isPure;
+        return result;
+    }
+};
+
+// 副作用分析结果类
+class SideEffectAnalysisResult : public AnalysisResultBase {
+private:
+    // 指令级别的副作用信息
+    std::unordered_map<Instruction*, SideEffectInfo> instructionSideEffects;
+    
+    // 函数级别的副作用信息
+    std::unordered_map<Function*, SideEffectInfo> functionSideEffects;
+    
+    // 已知的SysY标准库函数副作用信息
+    std::unordered_map<std::string, SideEffectInfo> knownFunctions;
+
+public:
+    SideEffectAnalysisResult();
+    virtual ~SideEffectAnalysisResult() noexcept override = default;
+    
+    // 获取指令的副作用信息
+    const SideEffectInfo& getInstructionSideEffect(Instruction* inst) const;
+    
+    // 获取函数的副作用信息
+    const SideEffectInfo& getFunctionSideEffect(Function* func) const;
+    
+    // 设置指令的副作用信息
+    void setInstructionSideEffect(Instruction* inst, const SideEffectInfo& info);
+    
+    // 设置函数的副作用信息
+    void setFunctionSideEffect(Function* func, const SideEffectInfo& info);
+    
+    // 检查指令是否有副作用
+    bool hasSideEffect(Instruction* inst) const;
+    
+    // 检查指令是否可能修改内存
+    bool mayModifyMemory(Instruction* inst) const;
+    
+    // 检查指令是否可能修改全局状态
+    bool mayModifyGlobal(Instruction* inst) const;
+    
+    // 检查函数是否为纯函数
+    bool isPureFunction(Function* func) const;
+    
+    // 获取已知函数的副作用信息
+    const SideEffectInfo* getKnownFunctionSideEffect(const std::string& funcName) const;
+    
+    // 初始化已知函数的副作用信息
+    void initializeKnownFunctions();
+    
+private:
+};
+
+// 副作用分析遍类 - Module级别分析
+class SysYSideEffectAnalysisPass : public AnalysisPass {
+public:
+    // 静态成员，作为该遍的唯一ID
+    static void* ID;
+    
+    SysYSideEffectAnalysisPass() : AnalysisPass("SysYSideEffectAnalysis", Granularity::Module) {}
+    
+    // 在模块上运行分析
+    bool runOnModule(Module* M, AnalysisManager& AM) override;
+    
+    // 获取分析结果
+    std::unique_ptr<AnalysisResultBase> getResult() override;
+    
+    // Pass 基类中的纯虚函数，必须实现
+    void* getPassID() const override { return &ID; }
+
+private:
+    // 分析结果
+    std::unique_ptr<SideEffectAnalysisResult> result;
+    
+    // 调用图分析结果
+    CallGraphAnalysisResult* callGraphAnalysis = nullptr;
+    
+    // 分析单个函数的副作用（Module级别的内部方法）
+    SideEffectInfo analyzeFunction(Function* func, AnalysisManager& AM);
+    
+    // 分析单个指令的副作用
+    SideEffectInfo analyzeInstruction(Instruction* inst, Function* currentFunc, AnalysisManager& AM);
+    
+    // 分析函数调用指令的副作用（利用调用图）
+    SideEffectInfo analyzeCallInstruction(CallInst* call, Function* currentFunc, AnalysisManager& AM);
+    
+    // 分析存储指令的副作用
+    SideEffectInfo analyzeStoreInstruction(StoreInst* store, Function* currentFunc, AnalysisManager& AM);
+    
+    // 分析内存设置指令的副作用
+    SideEffectInfo analyzeMemsetInstruction(MemsetInst* memset, Function* currentFunc, AnalysisManager& AM);
+    
+    // 使用不动点算法分析递归函数群
+    void analyzeStronglyConnectedComponent(const std::vector<Function*>& scc, AnalysisManager& AM);
+    
+    // 检查函数间副作用传播的收敛性
+    bool hasConverged(const std::unordered_map<Function*, SideEffectInfo>& oldEffects,
+                      const std::unordered_map<Function*, SideEffectInfo>& newEffects) const;
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/BuildCFG.h

@@ -0,0 +1,20 @@
+#pragma once
+
+#include "IR.h"
+#include "Pass.h"
+#include <queue>
+#include <set>
+
+namespace sysy {
+
+class BuildCFG : public OptimizationPass {
+public:
+  static void *ID;
+  BuildCFG() : OptimizationPass("BuildCFG", Granularity::Function) {}
+  bool runOnFunction(Function *F, AnalysisManager &AM) override;
+  void getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const override;
+  void *getPassID() const override { return &ID; }
+
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/DCE.h

@@ -4,6 +4,8 @@
 #include "IR.h"
 #include "SysYIROptUtils.h"
 #include "Dom.h" 
+#include "AliasAnalysis.h"
+#include "SideEffectAnalysis.h"
 #include <unordered_set>
 #include <queue>
 
@@ -25,8 +27,12 @@ public:
 private:
     // 存储活跃指令的集合
     std::unordered_set<Instruction*> alive_insts;
+    // 别名分析结果
+    AliasAnalysisResult* aliasAnalysis = nullptr;
+    // 副作用分析结果
+    SideEffectAnalysisResult* sideEffectAnalysis = nullptr;
 
-    // 判断指令是否是“天然活跃”的（即总是保留的）
+    // 判断指令是否是"天然活跃"的（即总是保留的）
     // inst: 要检查的指令
     // 返回值: 如果指令是天然活跃的，则为true，否则为false
     bool isAlive(Instruction* inst);
@@ -34,6 +40,9 @@ private:
     // 递归地将活跃指令及其依赖加入到 alive_insts 集合中
     // inst: 要标记为活跃的指令
     void addAlive(Instruction* inst);
+    
+    // 检查Store指令是否可能有副作用（通过别名分析）
+    bool mayHaveSideEffect(StoreInst* store);
 };
 
 // DCE 优化遍类，继承自 OptimizationPass

src/include/midend/Pass/Optimize/GVN.h

@@ -0,0 +1,87 @@
+#pragma once
+
+#include "Pass.h"
+#include "IR.h"
+#include "Dom.h"
+#include "SideEffectAnalysis.h"
+#include <unordered_map>
+#include <unordered_set>
+#include <vector>
+#include <string>
+#include <sstream>
+
+namespace sysy {
+
+// GVN优化遍的核心逻辑封装类
+class GVNContext {
+public:
+    // 运行GVN优化的主要方法
+    void run(Function* func, AnalysisManager* AM, bool& changed);
+
+private:
+    // 新的值编号系统
+    std::unordered_map<Value*, unsigned> valueToNumber;      // Value -> 值编号
+    std::unordered_map<unsigned, Value*> numberToValue;      // 值编号 -> 代表值
+    std::unordered_map<std::string, unsigned> expressionToNumber; // 表达式 -> 值编号
+    unsigned nextValueNumber = 1;
+    
+    // 已访问的基本块集合
+    std::unordered_set<BasicBlock*> visited;
+    
+    // 逆后序遍历的基本块列表
+    std::vector<BasicBlock*> rpoBlocks;
+    
+    // 需要删除的指令集合
+    std::unordered_set<Instruction*> needRemove;
+    
+    // 分析结果
+    DominatorTree* domTree = nullptr;
+    SideEffectAnalysisResult* sideEffectAnalysis = nullptr;
+    
+    // 计算逆后序遍历
+    void computeRPO(Function* func);
+    void dfs(BasicBlock* bb);
+    
+    // 新的值编号方法
+    unsigned getValueNumber(Value* value);
+    unsigned assignValueNumber(Value* value);
+    
+    // 基本块处理
+    void processBasicBlock(BasicBlock* bb, bool& changed);
+    
+    // 指令处理
+    bool processInstruction(Instruction* inst);
+    
+    // 表达式构建和查找
+    std::string buildExpressionKey(Instruction* inst);
+    Value* findExistingValue(const std::string& exprKey, Instruction* inst);
+    
+    // 支配关系和安全性检查
+    bool dominates(Instruction* a, Instruction* b);
+    bool isMemorySafe(LoadInst* earlierLoad, LoadInst* laterLoad);
+    
+    // 清理方法
+    void eliminateRedundantInstructions(bool& changed);
+    void invalidateMemoryValues(StoreInst* store);
+};
+
+// GVN优化遍类
+class GVN : public OptimizationPass {
+public:
+    // 静态成员，作为该遍的唯一ID
+    static void* ID;
+    
+    GVN() : OptimizationPass("GVN", Granularity::Function) {}
+    
+    // 在函数上运行优化
+    bool runOnFunction(Function* func, AnalysisManager& AM) override;
+    
+    // 返回该遍的唯一ID
+    void* getPassID() const override { return ID; }
+    
+    // 声明分析依赖
+    void getAnalysisUsage(std::set<void*>& analysisDependencies, 
+                         std::set<void*>& analysisInvalidations) const override;
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/GlobalStrengthReduction.h

@@ -0,0 +1,107 @@
+#pragma once
+
+#include "Pass.h"
+#include "IR.h"
+#include "SideEffectAnalysis.h"
+#include <unordered_map>
+#include <unordered_set>
+#include <vector>
+#include <cstdint>
+
+namespace sysy {
+
+// 魔数乘法结构，用于除法优化
+struct MagicNumber {
+    uint32_t multiplier;
+    int shift;
+    bool needAdd;
+    
+    MagicNumber(uint32_t m, int s, bool add = false) 
+        : multiplier(m), shift(s), needAdd(add) {}
+};
+
+// 全局强度削弱优化遍的核心逻辑封装类
+class GlobalStrengthReductionContext {
+public:
+    // 构造函数，接受IRBuilder参数
+    explicit GlobalStrengthReductionContext(IRBuilder* builder) : builder(builder) {}
+    
+    // 运行优化的主要方法
+    void run(Function* func, AnalysisManager* AM, bool& changed);
+
+private:
+    IRBuilder* builder;  // IR构建器
+    
+    // 分析结果
+    SideEffectAnalysisResult* sideEffectAnalysis = nullptr;
+    
+    // 优化计数
+    int algebraicOptCount = 0;
+    int strengthReductionCount = 0;
+    int divisionOptCount = 0;
+    
+    // 主要优化方法
+    bool processBasicBlock(BasicBlock* bb);
+    bool processInstruction(Instruction* inst);
+    
+    // 代数优化方法
+    bool tryAlgebraicOptimization(Instruction* inst);
+    bool optimizeAddition(BinaryInst* inst);
+    bool optimizeSubtraction(BinaryInst* inst);
+    bool optimizeMultiplication(BinaryInst* inst);
+    bool optimizeDivision(BinaryInst* inst);
+    bool optimizeComparison(BinaryInst* inst);
+    bool optimizeLogical(BinaryInst* inst);
+    
+    // 强度削弱方法
+    bool tryStrengthReduction(Instruction* inst);
+    bool reduceMultiplication(BinaryInst* inst);
+    bool reduceDivision(BinaryInst* inst);
+    bool reducePower(CallInst* inst);
+    
+    // 复杂乘法强度削弱方法
+    bool tryComplexMultiplication(BinaryInst* inst, Value* variable, int constant);
+    bool findOptimalShiftDecomposition(int constant, std::vector<int>& shifts);
+    Value* createShiftDecomposition(BinaryInst* inst, Value* variable, const std::vector<int>& shifts);
+    
+    // 魔数乘法相关方法
+    MagicNumber computeMagicNumber(uint32_t divisor);
+    std::pair<int, int> computeMulhMagicNumbers(int divisor);
+    Value* createMagicDivision(BinaryInst* divInst, uint32_t divisor, const MagicNumber& magic);
+    Value* createMagicDivisionLibdivide(BinaryInst* divInst, int divisor);
+    bool isPowerOfTwo(uint32_t n);
+    int log2OfPowerOfTwo(uint32_t n);
+    
+    // 辅助方法
+    bool isConstantInt(Value* val, int& constVal);
+    bool isConstantInt(Value* val, uint32_t& constVal);
+    ConstantInteger* getConstantInt(int val);
+    bool hasOnlyLocalUses(Instruction* inst);
+    void replaceWithOptimized(Instruction* original, Value* replacement);
+};
+
+// 全局强度削弱优化遍类
+class GlobalStrengthReduction : public OptimizationPass {
+private:
+    IRBuilder* builder;  // IR构建器，用于创建新指令
+
+public:
+    // 静态成员，作为该遍的唯一ID
+    static void* ID;
+    
+    // 构造函数，接受IRBuilder参数
+    explicit GlobalStrengthReduction(IRBuilder* builder) 
+        : OptimizationPass("GlobalStrengthReduction", Granularity::Function), builder(builder) {}
+    
+    // 在函数上运行优化
+    bool runOnFunction(Function* func, AnalysisManager& AM) override;
+    
+    // 返回该遍的唯一ID
+    void* getPassID() const override { return ID; }
+    
+    // 声明分析依赖
+    void getAnalysisUsage(std::set<void*>& analysisDependencies, 
+                         std::set<void*>& analysisInvalidations) const override;
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/InductionVariableElimination.h

@@ -0,0 +1,252 @@
+#pragma once
+
+#include "Pass.h"
+#include "IR.h"
+#include "LoopCharacteristics.h"
+#include "Loop.h"
+#include "Dom.h"
+#include "SideEffectAnalysis.h"
+#include "AliasAnalysis.h"
+#include <vector>
+#include <unordered_map>
+#include <unordered_set>
+#include <memory>
+
+namespace sysy {
+
+// 前向声明
+class LoopCharacteristicsResult;
+class LoopAnalysisResult;
+
+/**
+ * @brief 死归纳变量信息
+ * 记录一个可以被消除的归纳变量
+ */
+struct DeadInductionVariable {
+  PhiInst* phiInst;                    // phi 指令
+  std::vector<Instruction*> relatedInsts; // 相关的递增/递减指令
+  Loop* containingLoop;                // 所在循环
+  bool canEliminate;                   // 是否可以安全消除
+  
+  DeadInductionVariable(PhiInst* phi, Loop* loop) 
+    : phiInst(phi), containingLoop(loop), canEliminate(false) {}
+};
+
+/**
+ * @brief 归纳变量消除上下文类
+ * 封装归纳变量消除优化的核心逻辑和状态
+ */
+class InductionVariableEliminationContext {
+public:
+  InductionVariableEliminationContext() {}
+  
+  /**
+   * 运行归纳变量消除优化
+   * @param F 目标函数
+   * @param AM 分析管理器
+   * @return 是否修改了IR
+   */
+  bool run(Function* F, AnalysisManager& AM);
+
+private:
+  // 分析结果缓存
+  LoopAnalysisResult* loopAnalysis = nullptr;
+  LoopCharacteristicsResult* loopCharacteristics = nullptr;
+  DominatorTree* dominatorTree = nullptr;
+  SideEffectAnalysisResult* sideEffectAnalysis = nullptr;
+  AliasAnalysisResult* aliasAnalysis = nullptr;
+  
+  // 死归纳变量存储
+  std::vector<std::unique_ptr<DeadInductionVariable>> deadIVs;
+  std::unordered_map<Loop*, std::vector<DeadInductionVariable*>> loopToDeadIVs;
+  
+  // ========== 核心分析和优化阶段 ==========
+  
+  /**
+   * 阶段1：识别死归纳变量
+   * 找出没有被有效使用的归纳变量
+   */
+  void identifyDeadInductionVariables(Function* F);
+  
+  /**
+   * 阶段2：分析消除的安全性
+   * 确保消除操作不会破坏程序语义
+   */
+  void analyzeSafetyForElimination();
+  
+  /**
+   * 阶段3：执行归纳变量消除
+   * 删除死归纳变量及其相关指令
+   */
+  bool performInductionVariableElimination();
+  
+  // ========== 辅助方法 ==========
+  
+  /**
+   * 检查归纳变量是否为死归纳变量
+   * @param iv 归纳变量信息
+   * @param loop 所在循环
+   * @return 如果是死归纳变量返回相关信息，否则返回nullptr
+   */
+  std::unique_ptr<DeadInductionVariable> 
+  isDeadInductionVariable(const InductionVarInfo* iv, Loop* loop);
+  
+  /**
+   * 递归分析phi指令及其使用链是否都是死代码
+   * @param phiInst phi指令
+   * @param loop 所在循环
+   * @return phi指令是否可以安全删除
+   */
+  bool isPhiInstructionDeadRecursively(PhiInst* phiInst, Loop* loop);
+  
+  /**
+   * 递归分析指令的使用链是否都是死代码
+   * @param inst 要分析的指令
+   * @param loop 所在循环
+   * @param visited 已访问的指令集合（避免无限递归）
+   * @param currentPath 当前递归路径（检测循环依赖）
+   * @return 指令的使用链是否都是死代码
+   */
+  bool isInstructionUseChainDeadRecursively(Instruction* inst, Loop* loop, 
+                                           std::set<Instruction*>& visited, 
+                                           std::set<Instruction*>& currentPath);
+  
+  /**
+   * 检查循环是否有副作用
+   * @param loop 要检查的循环
+   * @return 循环是否有副作用
+   */
+  bool loopHasSideEffects(Loop* loop);
+  
+  /**
+   * 检查指令是否被用于循环退出条件
+   * @param inst 要检查的指令
+   * @param loop 所在循环
+   * @return 是否被用于循环退出条件
+   */
+  bool isUsedInLoopExitCondition(Instruction* inst, Loop* loop);
+  
+  /**
+   * 检查指令的结果是否未被有效使用
+   * @param inst 要检查的指令
+   * @param loop 所在循环
+   * @return 指令结果是否未被有效使用
+   */
+  bool isInstructionResultUnused(Instruction* inst, Loop* loop);
+  
+  /**
+   * 检查store指令是否存储到死地址（利用别名分析）
+   * @param store store指令
+   * @param loop 所在循环  
+   * @return 是否存储到死地址
+   */
+  bool isStoreToDeadLocation(StoreInst* store, Loop* loop);
+  
+  /**
+   * 检查指令是否为死代码或只在循环内部使用
+   * @param inst 要检查的指令
+   * @param loop 所在循环
+   * @return 是否为死代码或只在循环内部使用
+   */
+  bool isInstructionDeadOrInternalOnly(Instruction* inst, Loop* loop);
+  
+  /**
+   * 检查指令是否有效地为死代码（带递归深度限制）
+   * @param inst 要检查的指令
+   * @param loop 所在循环
+   * @param maxDepth 最大递归深度
+   * @return 指令是否有效地为死代码
+   */
+  bool isInstructionEffectivelyDead(Instruction* inst, Loop* loop, int maxDepth);
+  
+  /**
+   * 检查store指令是否有后续的load操作
+   * @param store store指令
+   * @param loop 所在循环
+   * @return 是否有后续的load操作
+   */
+  bool hasSubsequentLoad(StoreInst* store, Loop* loop);
+  
+  /**
+   * 检查指令是否在循环外有使用
+   * @param inst 要检查的指令
+   * @param loop 所在循环
+   * @return 是否在循环外有使用
+   */
+  bool hasUsageOutsideLoop(Instruction* inst, Loop* loop);
+  
+  /**
+   * 检查store指令是否在循环外有后续的load操作
+   * @param store store指令
+   * @param loop 所在循环
+   * @return 是否在循环外有后续的load操作
+   */
+  bool hasSubsequentLoadOutsideLoop(StoreInst* store, Loop* loop);
+  
+  /**
+   * 递归检查基本块子树中是否有对指定位置的load操作
+   * @param bb 基本块
+   * @param ptr 指针
+   * @param visited 已访问的基本块集合
+   * @return 是否有load操作
+   */
+  bool hasLoadInSubtree(BasicBlock* bb, Value* ptr, std::set<BasicBlock*>& visited);
+  
+  /**
+   * 收集与归纳变量相关的所有指令
+   * @param phiInst phi指令
+   * @param loop 所在循环
+   * @return 相关指令列表
+   */
+  std::vector<Instruction*> collectRelatedInstructions(PhiInst* phiInst, Loop* loop);
+  
+  /**
+   * 检查消除归纳变量的安全性
+   * @param deadIV 死归纳变量
+   * @return 是否可以安全消除
+   */
+  bool isSafeToEliminate(const DeadInductionVariable* deadIV);
+  
+  /**
+   * 消除单个死归纳变量
+   * @param deadIV 死归纳变量
+   * @return 是否成功消除
+   */
+  bool eliminateDeadInductionVariable(DeadInductionVariable* deadIV);
+  
+  /**
+   * 打印调试信息
+   */
+  void printDebugInfo();
+};
+
+/**
+ * @brief 归纳变量消除优化遍
+ * 消除循环中无用的归纳变量，减少寄存器压力
+ */
+class InductionVariableElimination : public OptimizationPass {
+public:
+  // 唯一的 Pass ID
+  static void *ID;
+  
+  InductionVariableElimination() 
+    : OptimizationPass("InductionVariableElimination", Granularity::Function) {}
+  
+  /**
+   * 在函数上运行归纳变量消除优化
+   * @param F 目标函数
+   * @param AM 分析管理器
+   * @return 是否修改了IR
+   */
+  bool runOnFunction(Function* F, AnalysisManager& AM) override;
+  
+  /**
+   * 声明分析依赖和失效信息
+   */
+  void getAnalysisUsage(std::set<void*>& analysisDependencies, 
+                       std::set<void*>& analysisInvalidations) const override;
+  
+  void* getPassID() const override { return &ID; }
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/LICM.h

@@ -0,0 +1,40 @@
+#pragma once
+#include "Pass.h"
+#include "Loop.h"
+#include "LoopCharacteristics.h"
+#include "Dom.h"
+#include <unordered_set>
+#include <vector>
+
+namespace sysy{
+
+class LICMContext {
+public:
+  LICMContext(Function* func, Loop* loop, IRBuilder* builder, const LoopCharacteristics* chars)
+    : func(func), loop(loop), builder(builder), chars(chars) {}
+    // 运行LICM主流程，返回IR是否被修改
+    bool run();
+
+private:
+    Function* func;
+    Loop* loop;
+    IRBuilder* builder;
+    const LoopCharacteristics* chars; // 特征分析结果
+
+    // 外提所有可提升指令
+    bool hoistInstructions();
+};
+
+
+class LICM : public OptimizationPass{
+private:
+	IRBuilder *builder; ///< IR构建器，用于插入指令
+public:
+	static void *ID;
+	LICM(IRBuilder *builder = nullptr) : OptimizationPass("LICM", Granularity::Function) , builder(builder) {}
+	bool runOnFunction(Function *F, AnalysisManager &AM) override;
+	void getAnalysisUsage(std::set<void *> &, std::set<void *> &) const override;
+	void *getPassID() const override { return &ID; }
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/LargeArrayToGlobal.h

@@ -0,0 +1,24 @@
+#pragma once
+
+#include "../Pass.h"
+
+namespace sysy {
+
+class LargeArrayToGlobalPass : public OptimizationPass {
+public:
+    static void *ID;
+
+    LargeArrayToGlobalPass() : OptimizationPass("LargeArrayToGlobal", Granularity::Module) {}
+
+    bool runOnModule(Module *M, AnalysisManager &AM) override;
+    void *getPassID() const override {
+        return &ID;
+    }
+
+private:
+    unsigned calculateTypeSize(Type *type);
+    void convertAllocaToGlobal(AllocaInst *alloca, Function *F, Module *M);
+    std::string generateUniqueGlobalName(AllocaInst *alloca, Function *F);
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/LoopNormalization.h

@@ -0,0 +1,155 @@
+#pragma once
+
+#include "Loop.h"          // 循环分析依赖
+#include "Dom.h"           // 支配树分析依赖
+#include "IR.h"            // IR定义
+#include "IRBuilder.h"     // IR构建器
+#include "Pass.h"          // Pass框架
+#include <memory>
+#include <set>
+#include <vector>
+
+namespace sysy {
+
+/**
+ * @brief 循环规范化转换Pass
+ * 
+ * 该Pass在循环不变量提升等优化前运行，主要负责：
+ * 1. 为没有前置块(preheader)的循环创建前置块
+ * 2. 确保循环结构符合后续优化的要求
+ * 3. 规范化循环的控制流结构
+ * 
+ * 前置块的作用：
+ * - 为循环不变量提升提供插入位置
+ * - 简化循环分析和优化
+ * - 确保循环有唯一的入口点
+ */
+class LoopNormalizationPass : public OptimizationPass {
+public:
+  // 唯一的 Pass ID
+  static void *ID;
+
+  LoopNormalizationPass(IRBuilder* builder) : OptimizationPass("LoopNormalization", Pass::Granularity::Function), builder(builder) {}
+
+  // 实现 getPassID
+  void *getPassID() const override { return &ID; }
+
+  // 核心运行方法
+  bool runOnFunction(Function *F, AnalysisManager &AM) override;
+  
+  // 声明分析依赖和失效信息
+  void getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const override;
+
+private:
+  // ========== IR构建器 ==========
+  IRBuilder* builder;                                  // IR构建器
+  
+  // ========== 缓存的分析结果 ==========
+  LoopAnalysisResult* loopAnalysis;                    // 循环结构分析结果
+  DominatorTree* domTree;                              // 支配树分析结果
+  
+  // ========== 规范化统计 ==========
+  struct NormalizationStats {
+    size_t totalLoops;                    // 总循环数
+    size_t loopsNeedingPreheader;        // 需要前置块的循环数
+    size_t preheadersCreated;            // 创建的前置块数
+    size_t loopsNormalized;              // 规范化的循环数
+    size_t redundantPhisRemoved;         // 删除的冗余PHI节点数
+    
+    NormalizationStats() : totalLoops(0), loopsNeedingPreheader(0), 
+                          preheadersCreated(0), loopsNormalized(0), 
+                          redundantPhisRemoved(0) {}
+  } stats;
+  
+  // ========== 核心规范化方法 ==========
+  
+  /**
+   * 规范化单个循环
+   * @param loop 要规范化的循环
+   * @return 是否进行了修改
+   */
+  bool normalizeLoop(Loop* loop);
+  
+  /**
+   * 为循环创建前置块
+   * @param loop 需要前置块的循环
+   * @return 创建的前置块，如果失败则返回nullptr
+   */
+  BasicBlock* createPreheaderForLoop(Loop* loop);
+  
+  /**
+   * 检查循环是否需要前置块（基于结构性需求）
+   * @param loop 要检查的循环
+   * @return true如果需要前置块
+   */
+  bool needsPreheader(Loop* loop);
+  
+  /**
+   * 检查循环是否已有合适的前置块
+   * @param loop 要检查的循环
+   * @return 现有的前置块，如果没有则返回nullptr
+   */
+  BasicBlock* getExistingPreheader(Loop* loop);
+  
+  /**
+   * 更新支配树关系（在创建新块后）
+   * @param newBlock 新创建的基本块
+   * @param loop 相关的循环
+   */
+  void updateDominatorRelations(BasicBlock* newBlock, Loop* loop);
+  
+  /**
+   * 重定向循环外的前驱块到新的前置块
+   * @param loop 目标循环
+   * @param preheader 新创建的前置块
+   * @param header 循环头部
+   */
+  void redirectExternalPredecessors(Loop* loop, BasicBlock* preheader, BasicBlock* header, const std::vector<BasicBlock*>& externalPreds);
+  
+  /**
+   * 为前置块生成合适的名称
+   * @param loop 相关的循环
+   * @return 生成的前置块名称
+   */
+  std::string generatePreheaderName(Loop* loop);
+  
+  /**
+   * 验证规范化结果的正确性
+   * @param loop 规范化后的循环
+   * @return true如果规范化正确
+   */
+  bool validateNormalization(Loop* loop);
+  
+  // ========== 辅助方法 ==========
+  
+  /**
+   * 获取循环的外部前驱块（不在循环内的前驱）
+   * @param loop 目标循环
+   * @return 外部前驱块列表
+   */
+  std::vector<BasicBlock*> getExternalPredecessors(Loop* loop);
+  
+  /**
+   * 检查基本块是否适合作为前置块
+   * @param block 候选基本块
+   * @param loop 目标循环
+   * @return true如果适合作为前置块
+   */
+  bool isSuitableAsPreheader(BasicBlock* block, Loop* loop);
+  
+  /**
+   * 更新PHI节点以适应新的前置块
+   * @param header 循环头部
+   * @param preheader 新的前置块
+   * @param oldPreds 原来的外部前驱
+   */
+  void updatePhiNodesForPreheader(BasicBlock* header, BasicBlock* preheader,
+                                 const std::vector<BasicBlock*>& oldPreds);
+  
+  /**
+   * 打印规范化统计信息
+   */
+  void printStats(Function* F);
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/LoopStrengthReduction.h

@@ -0,0 +1,233 @@
+#pragma once
+
+#include "Pass.h"
+#include "IR.h"
+#include "LoopCharacteristics.h"
+#include "Loop.h"
+#include "Dom.h"
+#include <vector>
+#include <unordered_map>
+#include <unordered_set>
+#include <memory>
+
+namespace sysy {
+
+// 前向声明
+class LoopCharacteristicsResult;
+class LoopAnalysisResult;
+
+/**
+ * @brief 强度削弱候选项信息
+ * 记录一个可以进行强度削弱的表达式信息
+ */
+struct StrengthReductionCandidate {
+  enum OpType {
+    MULTIPLY,    // 乘法: iv * const
+    DIVIDE,      // 除法: iv / 2^n (转换为右移)
+    DIVIDE_CONST, // 除法: iv / const (使用mulh指令优化)
+    REMAINDER    // 取模: iv % 2^n (转换为位与)
+  };
+  
+  enum DivisionStrategy {
+    SIMPLE_SHIFT,     // 简单右移（仅适用于无符号或非负数）
+    SIGNED_CORRECTION, // 有符号除法修正: (x + (x >> 31) & mask) >> k
+    MULH_OPTIMIZATION  // 使用mulh指令优化任意常数除法
+  };
+  
+  Instruction* originalInst;      // 原始指令 (如 i*4, i/8, i%16)
+  Value* inductionVar;            // 归纳变量 (如 i)
+  OpType operationType;           // 操作类型
+  DivisionStrategy divStrategy;   // 除法策略（仅用于除法）
+  int multiplier;                 // 乘数/除数/模数 (如 4, 8, 16)
+  int shiftAmount;                // 位移量 (对于2的幂)
+  int offset;                     // 偏移量 (如常数项)
+  BasicBlock* containingBlock;    // 所在基本块
+  Loop* containingLoop;           // 所在循环
+  bool hasNegativeValues;         // 归纳变量是否可能为负数
+  
+  // 强度削弱后的新变量
+  PhiInst* newPhi = nullptr;      // 新的 phi 指令
+  Value* newInductionVar = nullptr; // 新的归纳变量
+  
+  StrengthReductionCandidate(Instruction* inst, Value* iv, OpType opType, int value, int off, 
+                           BasicBlock* bb, Loop* loop)
+    : originalInst(inst), inductionVar(iv), operationType(opType), 
+      divStrategy(SIMPLE_SHIFT), multiplier(value), offset(off), 
+      containingBlock(bb), containingLoop(loop), hasNegativeValues(false) {
+    
+    // 计算位移量（用于除法和取模的强度削弱）
+    if (opType == DIVIDE || opType == REMAINDER) {
+      shiftAmount = 0;
+      int temp = value;
+      while (temp > 1) {
+        temp >>= 1;
+        shiftAmount++;
+      }
+    } else {
+      shiftAmount = 0;
+    }
+  }
+};
+
+/**
+ * @brief 强度削弱上下文类
+ * 封装强度削弱优化的核心逻辑和状态
+ */
+class StrengthReductionContext {
+public:
+  StrengthReductionContext(IRBuilder* builder) : builder(builder) {}
+  
+  /**
+   * 运行强度削弱优化
+   * @param F 目标函数
+   * @param AM 分析管理器
+   * @return 是否修改了IR
+   */
+  bool run(Function* F, AnalysisManager& AM);
+
+private:
+  IRBuilder* builder;
+  
+  // 分析结果缓存
+  LoopAnalysisResult* loopAnalysis = nullptr;
+  LoopCharacteristicsResult* loopCharacteristics = nullptr;
+  DominatorTree* dominatorTree = nullptr;
+  
+  // 候选项存储
+  std::vector<std::unique_ptr<StrengthReductionCandidate>> candidates;
+  std::unordered_map<Loop*, std::vector<StrengthReductionCandidate*>> loopToCandidates;
+  
+  // ========== 核心分析和优化阶段 ==========
+  
+  /**
+   * 阶段1：识别强度削弱候选项
+   * 扫描所有循环中的乘法指令，找出可以优化的模式
+   */
+  void identifyStrengthReductionCandidates(Function* F);
+  
+  /**
+   * 阶段2：分析候选项的优化潜力
+   * 评估每个候选项的收益，过滤掉不值得优化的情况
+   */
+  void analyzeOptimizationPotential();
+  
+  /**
+   * 阶段3：执行强度削弱变换
+   * 对选中的候选项执行实际的强度削弱优化
+   */
+  bool performStrengthReduction();
+  
+  // ========== 辅助分析函数 ==========
+  
+  /**
+   * 分析归纳变量是否可能取负值
+   * @param ivInfo 归纳变量信息
+   * @param loop 所属循环
+   * @return 如果可能为负数返回true
+   */
+  bool analyzeInductionVariableRange(const InductionVarInfo* ivInfo, Loop* loop) const;
+  
+  /**
+   * 生成除法替换代码
+   * @param candidate 优化候选项
+   * @param builder IR构建器
+   * @return 替换值
+   */
+  Value* generateDivisionReplacement(StrengthReductionCandidate* candidate, IRBuilder* builder) const;
+  
+  /**
+   * 生成任意常数除法替换代码
+   * @param candidate 优化候选项
+   * @param builder IR构建器
+   * @return 替换值
+   */
+  Value* generateConstantDivisionReplacement(StrengthReductionCandidate* candidate, IRBuilder* builder) const;
+  
+  /**
+   * 检查指令是否为强度削弱候选项
+   * @param inst 要检查的指令
+   * @param loop 所在循环
+   * @return 如果是候选项返回候选项信息，否则返回nullptr
+   */
+  std::unique_ptr<StrengthReductionCandidate> 
+  isStrengthReductionCandidate(Instruction* inst, Loop* loop);
+  
+  /**
+   * 检查值是否为循环的归纳变量
+   * @param val 要检查的值
+   * @param loop 循环
+   * @param characteristics 循环特征信息
+   * @return 如果是归纳变量返回归纳变量信息，否则返回nullptr
+   */
+  const InductionVarInfo* 
+  getInductionVarInfo(Value* val, Loop* loop, const LoopCharacteristics* characteristics);
+  
+  /**
+   * 为候选项创建新的归纳变量
+   * @param candidate 候选项
+   * @return 是否成功创建
+   */
+  bool createNewInductionVariable(StrengthReductionCandidate* candidate);
+  
+  /**
+   * 替换原始指令的所有使用
+   * @param candidate 候选项
+   * @return 是否成功替换
+   */
+  bool replaceOriginalInstruction(StrengthReductionCandidate* candidate);
+  
+  /**
+   * 估算优化收益
+   * 计算强度削弱后的性能提升
+   * @param candidate 候选项
+   * @return 估算的收益分数
+   */
+  double estimateOptimizationBenefit(const StrengthReductionCandidate* candidate);
+  
+  /**
+   * 检查优化的合法性
+   * @param candidate 候选项
+   * @return 是否可以安全地进行优化
+   */
+  bool isOptimizationLegal(const StrengthReductionCandidate* candidate);
+  
+  /**
+   * 打印调试信息
+   */
+  void printDebugInfo();
+};
+
+/**
+ * @brief 循环强度削弱优化遍
+ * 将循环中的乘法运算转换为更高效的加法运算
+ */
+class LoopStrengthReduction : public OptimizationPass {
+public:
+  // 唯一的 Pass ID
+  static void *ID;
+  
+  LoopStrengthReduction(IRBuilder* builder) 
+    : OptimizationPass("LoopStrengthReduction", Granularity::Function), 
+      builder(builder) {}
+  
+  /**
+   * 在函数上运行强度削弱优化
+   * @param F 目标函数
+   * @param AM 分析管理器
+   * @return 是否修改了IR
+   */
+  bool runOnFunction(Function* F, AnalysisManager& AM) override;
+  
+  /**
+   * 声明分析依赖和失效信息
+   */
+  void getAnalysisUsage(std::set<void*>& analysisDependencies, 
+                       std::set<void*>& analysisInvalidations) const override;
+  
+  void* getPassID() const override { return &ID; }
+
+private:
+  IRBuilder* builder;
+};
+
+} // namespace sysy

src/include/midend/Pass/Optimize/SCCP.h

@@ -3,6 +3,8 @@
 #include "IR.h"
 #include "Pass.h"
 #include "SysYIROptUtils.h"
+#include "AliasAnalysis.h"
+#include "SideEffectAnalysis.h"
 #include <cassert>
 #include <iostream>
 #include <map>
@@ -63,6 +65,8 @@ struct SSAPValue {
 class SCCPContext {
 private:
   IRBuilder *builder; // IR 构建器，用于插入指令和创建常量
+  AliasAnalysisResult *aliasAnalysis; // 别名分析结果
+  SideEffectAnalysisResult *sideEffectAnalysis; // 副作用分析结果
 
   // 工作列表
   // 存储需要重新评估的指令
@@ -92,6 +96,14 @@ private:
   SSAPValue ComputeConstant(BinaryInst *binaryinst, SSAPValue lhsVal, SSAPValue rhsVal);
   // 辅助函数：对一元操作进行常量折叠
   SSAPValue ComputeConstant(UnaryInst *unaryInst, SSAPValue operandVal);
+  // 辅助函数：检查是否为已知的纯函数
+  bool isKnownPureFunction(const std::string &funcName) const;
+  // 辅助函数：计算纯函数的常量结果
+  SSAPValue computePureFunctionResult(CallInst *call, const std::vector<SSAPValue> &argValues);
+  // 辅助函数：查找存储到指定位置的常量值
+  SSAPValue findStoredConstantValue(Value *ptr, BasicBlock *currentBB);
+  // 辅助函数：动态检查数组访问是否为常量索引（考虑SCCP状态）
+  bool hasRuntimeConstantAccess(Value *ptr);
 
   // 主要优化阶段
   // 阶段1: 常量传播与折叠
@@ -117,7 +129,13 @@ private:
   void RemovePhiIncoming(BasicBlock *phiParentBB, BasicBlock *removedPred);
 
 public:
-  SCCPContext(IRBuilder *builder) : builder(builder) {}
+  SCCPContext(IRBuilder *builder) : builder(builder), aliasAnalysis(nullptr), sideEffectAnalysis(nullptr) {}
+  
+  // 设置别名分析结果
+  void setAliasAnalysis(AliasAnalysisResult *aa) { aliasAnalysis = aa; }
+  
+  // 设置副作用分析结果
+  void setSideEffectAnalysis(SideEffectAnalysisResult *sea) { sideEffectAnalysis = sea; }
 
   // 运行 SCCP 优化
   void run(Function *func, AnalysisManager &AM);

src/include/midend/Pass/Optimize/SysYIROptUtils.h

@@ -48,13 +48,6 @@ public:
             }
         }
     }
-    // 清空 User 的 operands 向量。这会递减 User 持有的 shared_ptr<Use> 的引用计数。
-    // 当引用计数降为 0 时，Use 对象本身将被销毁。
-    // User::operands.clear(); // 这个步骤会在 Instruction 的析构函数中自动完成，因为它是 vector 成员
-                             // 或者我们可以在 User::removeOperand 方法中确保 Use 对象从 operands 中移除。
-                             // 实际上，只要 Value::removeUse(use_ptr) 被调用了，
-                             // 当 Instruction 所在的 unique_ptr 销毁时，它的 operands vector 也会被销毁。
-                             // 所以这里不需要显式 clear()
   }
   static void usedelete(Instruction *inst) {
     assert(inst && "Instruction to delete cannot be null.");
@@ -75,7 +68,7 @@ public:
     // 步骤3: 物理删除指令
     // 这会导致 Instruction 对象的 unique_ptr 销毁，从而调用其析构函数链。
     parentBlock->removeInst(inst);
-}
+} 
 
   static BasicBlock::iterator usedelete(BasicBlock::iterator inst_it) {
     Instruction *inst_to_delete = inst_it->get();
@@ -92,7 +85,7 @@ public:
 
     // 步骤3: 物理删除指令并返回下一个迭代器
     return parentBlock->removeInst(inst_it);
-}
+  }
 
   // 判断是否是全局变量
   static bool isGlobal(Value *val) {
@@ -114,6 +107,190 @@ public:
     // 所以当AllocaInst的basetype是PointerType时（一维数组）或者是指向ArrayType的PointerType（多位数组）时，返回true
     return aval && (baseType->isPointer() || baseType->as<PointerType>()->getBaseType()->isArray());
   }
+
+  
+  //该实现参考了libdivide的算法
+  static std::pair<int, int> computeMulhMagicNumbers(int divisor) {
+    
+    if (DEBUG) {
+      std::cout << "\n[SR] ===== Computing magic numbers for divisor " << divisor << " (libdivide algorithm) =====" << std::endl;
+    }
+    
+    if (divisor == 0) {
+      if (DEBUG) std::cout << "[SR] Error: divisor must be != 0" << std::endl;
+      return {-1, -1};
+    }
+
+    // libdivide 常数
+    const uint8_t LIBDIVIDE_ADD_MARKER = 0x40;
+    const uint8_t LIBDIVIDE_NEGATIVE_DIVISOR = 0x80;
+    
+    // 辅助函数：计算前导零个数
+    auto count_leading_zeros32 = [](uint32_t val) -> uint32_t {
+      if (val == 0) return 32;
+      return __builtin_clz(val);
+    };
+    
+    // 辅助函数：64位除法返回32位商和余数
+    auto div_64_32 = [](uint32_t high, uint32_t low, uint32_t divisor, uint32_t* rem) -> uint32_t {
+      uint64_t dividend = ((uint64_t)high << 32) | low;
+      uint32_t quotient = dividend / divisor;
+      *rem = dividend % divisor;
+      return quotient;
+    };
+
+    if (DEBUG) {
+      std::cout << "[SR] Input divisor: " << divisor << std::endl;
+    }
+
+    // libdivide_internal_s32_gen 算法实现
+    int32_t d = divisor;
+    uint32_t ud = (uint32_t)d;
+    uint32_t absD = (d < 0) ? -ud : ud;
+    
+    if (DEBUG) {
+      std::cout << "[SR] absD = " << absD << std::endl;
+    }
+    
+    uint32_t floor_log_2_d = 31 - count_leading_zeros32(absD);
+    
+    if (DEBUG) {
+      std::cout << "[SR] floor_log_2_d = " << floor_log_2_d << std::endl;
+    }
+    
+    // 检查 absD 是否为2的幂
+    if ((absD & (absD - 1)) == 0) {
+      if (DEBUG) {
+        std::cout << "[SR] " << absD << " 是2的幂，使用移位方法" << std::endl;
+      }
+      
+      // 对于2的幂，我们只使用移位，不需要魔数
+      int shift = floor_log_2_d;
+      if (d < 0) shift |= 0x80; // 标记负数
+      
+      if (DEBUG) {
+        std::cout << "[SR] Power of 2 result: magic=0, shift=" << shift << std::endl;
+        std::cout << "[SR] ===== End magic computation =====" << std::endl;
+      }
+      
+      // 对于我们的目的，我们将在IR生成中以不同方式处理2的幂
+      // 返回特殊标记
+      return {0, shift};
+    }
+    
+    if (DEBUG) {
+      std::cout << "[SR] " << absD << " is not a power of 2, computing magic number" << std::endl;
+    }
+    
+    // 非2的幂除数的魔数计算
+    uint8_t more;
+    uint32_t rem, proposed_m;
+    
+    // 计算 proposed_m = floor(2^(floor_log_2_d + 31) / absD)
+    proposed_m = div_64_32((uint32_t)1 << (floor_log_2_d - 1), 0, absD, &rem);
+    const uint32_t e = absD - rem;
+    
+    if (DEBUG) {
+      std::cout << "[SR] proposed_m = " << proposed_m << ", rem = " << rem << ", e = " << e << std::endl;
+    }
+    
+    // 确定是否需要"加法"版本
+    const bool branchfree = false; // 使用分支版本
+    
+    if (!branchfree && e < ((uint32_t)1 << floor_log_2_d)) {
+      // 这个幂次有效
+      more = (uint8_t)(floor_log_2_d - 1);
+      if (DEBUG) {
+        std::cout << "[SR] Using basic algorithm, shift = " << (int)more << std::endl;
+      }
+    } else {
+      // 我们需要上升一个等级
+      proposed_m += proposed_m;
+      const uint32_t twice_rem = rem + rem;
+      if (twice_rem >= absD || twice_rem < rem) {
+        proposed_m += 1;
+      }
+      more = (uint8_t)(floor_log_2_d | LIBDIVIDE_ADD_MARKER);
+      if (DEBUG) {
+        std::cout << "[SR] Using add algorithm, proposed_m = " << proposed_m << ", more = " << (int)more << std::endl;
+      }
+    }
+    
+    proposed_m += 1;
+    int32_t magic = (int32_t)proposed_m;
+    
+    // 处理负除数
+    if (d < 0) {
+      more |= LIBDIVIDE_NEGATIVE_DIVISOR;
+      if (!branchfree) {
+        magic = -magic;
+      }
+      if (DEBUG) {
+        std::cout << "[SR] Negative divisor, magic = " << magic << ", more = " << (int)more << std::endl;
+      }
+    }
+    
+    // 为我们的IR生成提取移位量和标志  
+    int shift = more & 0x3F; // 移除标志，保留移位量（位0-5）
+    bool need_add = (more & LIBDIVIDE_ADD_MARKER) != 0;
+    bool is_negative = (more & LIBDIVIDE_NEGATIVE_DIVISOR) != 0;
+    
+    if (DEBUG) {
+      std::cout << "[SR] Final result: magic = " << magic << ", more = " << (int)more 
+                << " (0x" << std::hex << (int)more << std::dec << ")" << std::endl;
+      std::cout << "[SR] Shift = " << shift << ", need_add = " << need_add 
+                << ", is_negative = " << is_negative << std::endl;
+      
+      // Test the magic number using the correct libdivide algorithm
+      std::cout << "[SR] Testing magic number (libdivide algorithm):" << std::endl;
+      int test_values[] = {1, 7, 37, 100, 999, -1, -7, -37, -100};
+      
+      for (int test_val : test_values) {
+        int64_t quotient;
+        
+        // 实现正确的libdivide算法
+        int64_t product = (int64_t)test_val * magic;
+        int64_t high_bits = product >> 32;
+        
+        if (need_add) {
+          // ADD_MARKER情况：移位前加上被除数
+          // 这是libdivide的关键洞察！
+          high_bits += test_val;
+          quotient = high_bits >> shift;
+        } else {
+          // 正常情况：只是移位
+          quotient = high_bits >> shift;
+        }
+        
+        // 符号修正：这是libdivide有符号除法的关键部分！
+        // 如果被除数为负，商需要加1来匹配C语言的截断除法语义
+        if (test_val < 0) {
+          quotient += 1;
+        }
+        
+        int expected = test_val / divisor;
+        
+        bool correct = (quotient == expected);
+        std::cout << "[SR]   " << test_val << " / " << divisor << " = " << quotient 
+                  << " (expected " << expected << ") " << (correct ? "✓" : "✗") << std::endl;
+      }
+      
+      std::cout << "[SR] ===== End magic computation =====" << std::endl;
+    }
+    
+    // 返回魔数、移位量，并在移位中编码ADD_MARKER标志
+    // 我们将使用移位的第6位表示ADD_MARKER，第7位表示负数（如果需要）
+    int encoded_shift = shift;
+    if (need_add) {
+      encoded_shift |= 0x40; // 设置第6位表示ADD_MARKER
+      if (DEBUG) {
+        std::cout << "[SR] Encoding ADD_MARKER in shift: " << encoded_shift << std::endl;
+      }
+    }
+    
+    return {magic, encoded_shift};
+  }
+
 };
 
 }// namespace sysy

src/include/midend/Pass/Optimize/TailCallOpt.h

@@ -0,0 +1,39 @@
+#pragma once
+
+#include "Pass.h"
+#include "Dom.h"
+#include "Loop.h"
+
+namespace sysy {
+
+/**
+ * @class TailCallOpt
+ * @brief 优化尾调用的中端优化通道。
+ *
+ * 该类实现了一个针对函数级别的尾调用优化的优化通道（OptimizationPass）。
+ * 通过分析和转换 IR（中间表示），将可优化的尾调用转换为更高效的形式，
+ * 以减少函数调用的开销，提升程序性能。
+ *
+ * @note 需要传入 IRBuilder 指针用于 IR 构建和修改。
+ *
+ * @method runOnFunction
+ * 对指定函数进行尾调用优化。
+ *
+ * @method getPassID
+ * 获取当前优化通道的唯一标识符。
+ *
+ * @method getAnalysisUsage
+ * 指定该优化通道所依赖和失效的分析集合。
+ */
+class TailCallOpt : public OptimizationPass {
+private:
+  IRBuilder* builder;
+public:
+  TailCallOpt(IRBuilder* builder) : OptimizationPass("TailCallOpt", Granularity::Function), builder(builder) {}
+  static void *ID;
+  bool runOnFunction(Function *F, AnalysisManager &AM) override;
+  void *getPassID() const override { return &ID; }
+  void getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const override;
+};
+
+} // namespace sysy

src/include/midend/Pass/Pass.h

@@ -151,17 +151,21 @@ public:
     }
     AnalysisPass *analysisPass = static_cast<AnalysisPass *>(basePass.get());
 
-    if(DEBUG){
-      std::cout << "Running Analysis Pass: " << analysisPass->getName() << "\n";
-    }
     // 根据分析遍的粒度处理
     switch (analysisPass->getGranularity()) {
       case Pass::Granularity::Module: {
         // 检查是否已存在有效结果
         auto it = moduleCachedResults.find(analysisID);
         if (it != moduleCachedResults.end()) {
+          if(DEBUG) {
+            std::cout << "Using cached result for Analysis Pass: " << analysisPass->getName() << "\n";
+          }
           return static_cast<T *>(it->second.get()); // 返回缓存结果
         }
+        // 只有在实际运行时才打印调试信息
+        if(DEBUG){
+          std::cout << "Running Analysis Pass: " << analysisPass->getName() << "\n";
+        }
         // 运行模块级分析遍
         if (!pModuleRef) {
           std::cerr << "Error: Module reference not set for AnalysisManager to run Module Pass.\n";
@@ -183,8 +187,16 @@ public:
         // 检查是否已存在有效结果
         auto it = functionCachedResults.find({F, analysisID});
         if (it != functionCachedResults.end()) {
+          if(DEBUG) {
+            std::cout << "Using cached result for Analysis Pass: " << analysisPass->getName() << " (Function: " << F->getName() << ")\n";
+          }
           return static_cast<T *>(it->second.get()); // 返回缓存结果
         }
+        // 只有在实际运行时才打印调试信息
+        if(DEBUG){
+          std::cout << "Running Analysis Pass: " << analysisPass->getName() << "\n";
+          std::cout << "Function: " << F->getName() << "\n";
+        }
         // 运行函数级分析遍
         analysisPass->runOnFunction(F, *this);
         // 获取结果并缓存
@@ -202,8 +214,16 @@ public:
         // 检查是否已存在有效结果
         auto it = basicBlockCachedResults.find({BB, analysisID});
         if (it != basicBlockCachedResults.end()) {
+          if(DEBUG) {
+            std::cout << "Using cached result for Analysis Pass: " << analysisPass->getName() << " (BasicBlock: " << BB->getName() << ")\n";
+          }
           return static_cast<T *>(it->second.get()); // 返回缓存结果
         }
+        // 只有在实际运行时才打印调试信息
+        if(DEBUG){
+          std::cout << "Running Analysis Pass: " << analysisPass->getName() << "\n";
+          std::cout << "BasicBlock: " << BB->getName() << "\n";
+        }
         // 运行基本块级分析遍
         analysisPass->runOnBasicBlock(BB, *this);
         // 获取结果并缓存
@@ -279,7 +299,7 @@ private:
   IRBuilder *pBuilder;
 
 public:
-  PassManager() = default;
+  PassManager() = delete;
   ~PassManager() = default;
 
   PassManager(Module *module, IRBuilder *builder) : pmodule(module) ,pBuilder(builder), analysisManager(module) {}

src/include/midend/SysYIRGenerator.h

@@ -86,7 +86,60 @@ private:
     case LPAREN: case RPAREN: return 0; // Parentheses have lowest precedence for stack logic
     default: return -1; // Unknown operator
     }
-  }
+  };
+
+  struct ExpKey {
+    BinaryOp op;  ///< 操作符
+    Value *left;  ///< 左操作数
+    Value *right; ///< 右操作数
+    ExpKey(BinaryOp op, Value *left, Value *right) : op(op), left(left), right(right) {}
+
+    bool operator<(const ExpKey &other) const {
+      if (op != other.op)
+        return op < other.op; ///< 比较操作符
+      if (left != other.left)
+        return left < other.left; ///< 比较左操作
+      return right < other.right; ///< 比较右操作数
+    } ///< 重载小于运算符用于比较ExpKey
+  };
+
+  struct UnExpKey {
+    BinaryOp op;    ///< 一元操作符
+    Value *operand; ///< 操作数
+    UnExpKey(BinaryOp op, Value *operand) : op(op), operand(operand) {}
+
+    bool operator<(const UnExpKey &other) const {
+      if (op != other.op)
+        return op < other.op;         ///< 比较操作符
+      return operand < other.operand; ///< 比较操作数
+    } ///< 重载小于运算符用于比较UnExpKey
+  };
+
+  struct GEPKey {
+    Value *basePointer;
+    std::vector<Value *> indices;
+
+    // 为 std::map 定义比较运算符，使得 GEPKey 可以作为键
+    bool operator<(const GEPKey &other) const {
+      if (basePointer != other.basePointer) {
+        return basePointer < other.basePointer;
+      }
+      // 逐个比较索引，确保顺序一致
+      if (indices.size() != other.indices.size()) {
+        return indices.size() < other.indices.size();
+      }
+      for (size_t i = 0; i < indices.size(); ++i) {
+        if (indices[i] != other.indices[i]) {
+          return indices[i] < other.indices[i];
+        }
+      }
+      return false; // 如果 basePointer 和所有索引都相同，则认为相等
+    }
+  };
+  std::map<GEPKey, Value*> availableGEPs; ///< 用于存储 GEP 的缓存
+  std::map<ExpKey, Value*> availableBinaryExpressions;
+  std::map<UnExpKey, Value*> availableUnaryExpressions;
+  std::map<Value*, Value*> availableLoads;
 
 public:
   SysYIRGenerator() = default;
@@ -167,6 +220,15 @@ public:
   Value* computeExp(SysYParser::ExpContext *ctx, Type* targetType = nullptr);
   Value* computeAddExp(SysYParser::AddExpContext *ctx, Type* targetType = nullptr);
   void compute();
+
+  // 参数是发生 store 操作的目标地址/变量的 Value*
+  void invalidateExpressionsOnStore(Value* storedAddress);
+
+  // 清除因函数调用而失效的表达式缓存（保守策略）
+  void invalidateExpressionsOnCall();
+
+  // 在进入新的基本块时清空所有表达式缓存
+  void enterNewBasicBlock();
 public:
   // 获取GEP指令的地址
   Value* getGEPAddressInst(Value* basePointer, const std::vector<Value*>& indices);

src/midend/CMakeLists.txt

@@ -6,11 +6,26 @@ add_library(midend_lib STATIC
     Pass/Pass.cpp
     Pass/Analysis/Dom.cpp
     Pass/Analysis/Liveness.cpp
+    Pass/Analysis/Loop.cpp
+    Pass/Analysis/LoopCharacteristics.cpp
+    Pass/Analysis/LoopVectorization.cpp
+    Pass/Analysis/AliasAnalysis.cpp
+    Pass/Analysis/SideEffectAnalysis.cpp
+    Pass/Analysis/CallGraphAnalysis.cpp
     Pass/Optimize/DCE.cpp
     Pass/Optimize/Mem2Reg.cpp
     Pass/Optimize/Reg2Mem.cpp
+    Pass/Optimize/GVN.cpp
     Pass/Optimize/SysYIRCFGOpt.cpp
     Pass/Optimize/SCCP.cpp
+    Pass/Optimize/LoopNormalization.cpp
+    Pass/Optimize/LICM.cpp
+    Pass/Optimize/LoopStrengthReduction.cpp
+    Pass/Optimize/InductionVariableElimination.cpp
+    Pass/Optimize/GlobalStrengthReduction.cpp
+    Pass/Optimize/BuildCFG.cpp
+    Pass/Optimize/LargeArrayToGlobal.cpp
+    Pass/Optimize/TailCallOpt.cpp
 )
 
 # 包含中端模块所需的头文件路径

src/midend/Pass/Analysis/AliasAnalysis.cpp

@@ -0,0 +1,559 @@
+#include "AliasAnalysis.h"
+#include "SysYIRPrinter.h"
+#include <iostream>
+
+extern int DEBUG;
+
+namespace sysy {
+
+// 静态成员初始化
+void *SysYAliasAnalysisPass::ID = (void *)&SysYAliasAnalysisPass::ID;
+
+// ========== AliasAnalysisResult 实现 ==========
+
+void AliasAnalysisResult::print() const {
+    std::cout << "---- Alias Analysis Results for Function: " << AssociatedFunction->getName() << " ----\n";
+    
+    // 打印内存位置信息
+    std::cout << "  Memory Locations (" << LocationMap.size() << "):\n";
+    for (const auto& pair : LocationMap) {
+        const auto& loc = pair.second;
+        std::cout << "    - Base: " << loc->basePointer->getName();
+        std::cout << " (Type: ";
+        if (loc->isLocalArray) std::cout << "Local";
+        else if (loc->isFunctionParameter) std::cout << "Parameter";
+        else if (loc->isGlobalArray) std::cout << "Global";
+        else std::cout << "Unknown";
+        std::cout << ")\n";
+    }
+    
+    // 打印别名关系
+    std::cout << "  Alias Relations (" << AliasMap.size() << "):\n";
+    for (const auto& pair : AliasMap) {
+        std::cout << "    - (" << pair.first.first->getName() << ", " << pair.first.second->getName() << "): ";
+        switch (pair.second) {
+            case AliasType::NO_ALIAS: std::cout << "No Alias"; break;
+            case AliasType::SELF_ALIAS: std::cout << "Self Alias"; break;
+            case AliasType::POSSIBLE_ALIAS: std::cout << "Possible Alias"; break;
+            case AliasType::UNKNOWN_ALIAS: std::cout << "Unknown Alias"; break;
+        }
+        std::cout << "\n";
+    }
+    std::cout << "-----------------------------------------------------------\n";
+}
+
+AliasType AliasAnalysisResult::queryAlias(Value* ptr1, Value* ptr2) const {
+  auto key = std::make_pair(ptr1, ptr2);
+  auto it = AliasMap.find(key);
+  if (it != AliasMap.end()) {
+    return it->second;
+  }
+  
+  // 尝试反向查找
+  key = std::make_pair(ptr2, ptr1);
+  it = AliasMap.find(key);
+  if (it != AliasMap.end()) {
+    return it->second;
+  }
+  
+  return AliasType::UNKNOWN_ALIAS; // 保守估计
+}
+
+const MemoryLocation* AliasAnalysisResult::getMemoryLocation(Value* ptr) const {
+  auto it = LocationMap.find(ptr);
+  return (it != LocationMap.end()) ? it->second.get() : nullptr;
+}
+
+bool AliasAnalysisResult::isLocalArray(Value* ptr) const {
+  const MemoryLocation* loc = getMemoryLocation(ptr);
+  return loc && loc->isLocalArray;
+}
+
+bool AliasAnalysisResult::isFunctionParameter(Value* ptr) const {
+  const MemoryLocation* loc = getMemoryLocation(ptr);
+  return loc && loc->isFunctionParameter;
+}
+
+bool AliasAnalysisResult::isGlobalArray(Value* ptr) const {
+  const MemoryLocation* loc = getMemoryLocation(ptr);
+  return loc && loc->isGlobalArray;
+}
+
+bool AliasAnalysisResult::hasConstantAccess(Value* ptr) const {
+  const MemoryLocation* loc = getMemoryLocation(ptr);
+  return loc && loc->hasConstantIndices;
+}
+
+AliasAnalysisResult::Statistics AliasAnalysisResult::getStatistics() const {
+  Statistics stats = {0};
+  
+  stats.totalQueries = AliasMap.size();
+  
+  for (auto& pair : AliasMap) {
+    switch (pair.second) {
+      case AliasType::NO_ALIAS: stats.noAlias++; break;
+      case AliasType::SELF_ALIAS: stats.selfAlias++; break;
+      case AliasType::POSSIBLE_ALIAS: stats.possibleAlias++; break;
+      case AliasType::UNKNOWN_ALIAS: stats.unknownAlias++; break;
+    }
+  }
+  
+  for (auto& loc : LocationMap) {
+    if (loc.second->isLocalArray) stats.localArrays++;
+    if (loc.second->isFunctionParameter) stats.functionParameters++;
+    if (loc.second->isGlobalArray) stats.globalArrays++;
+    if (loc.second->hasConstantIndices) stats.constantAccesses++;
+  }
+  
+  return stats;
+}
+
+void AliasAnalysisResult::printStatics() const {
+  std::cout << "=== Alias Analysis Results ===" << std::endl;
+  
+  auto stats = getStatistics();
+  std::cout << "Total queries: " << stats.totalQueries << std::endl;
+  std::cout << "No alias: " << stats.noAlias << std::endl;
+  std::cout << "Self alias: " << stats.selfAlias << std::endl;
+  std::cout << "Possible alias: " << stats.possibleAlias << std::endl;
+  std::cout << "Unknown alias: " << stats.unknownAlias << std::endl;
+  std::cout << "Local arrays: " << stats.localArrays << std::endl;
+  std::cout << "Function parameters: " << stats.functionParameters << std::endl;
+  std::cout << "Global arrays: " << stats.globalArrays << std::endl;
+  std::cout << "Constant accesses: " << stats.constantAccesses << std::endl;
+}
+
+void AliasAnalysisResult::addMemoryLocation(std::unique_ptr<MemoryLocation> location) {
+  Value* ptr = location->accessPointer;
+  LocationMap[ptr] = std::move(location);
+}
+
+void AliasAnalysisResult::addAliasRelation(Value* ptr1, Value* ptr2, AliasType type) {
+  auto key = std::make_pair(ptr1, ptr2);
+  AliasMap[key] = type;
+}
+
+// ========== SysYAliasAnalysisPass 实现 ==========
+
+bool SysYAliasAnalysisPass::runOnFunction(Function *F, AnalysisManager &AM) {
+  if (DEBUG) {
+    std::cout << "Running SysY Alias Analysis on function: " << F->getName() << std::endl;
+  }
+  
+  // 创建分析结果
+  CurrentResult = std::make_unique<AliasAnalysisResult>(F);
+  
+  // 执行主要分析步骤
+  collectMemoryAccesses(F);
+  buildAliasRelations(F);
+  optimizeForSysY(F);
+  
+  if (DEBUG) {
+    CurrentResult->print();
+    CurrentResult->printStatics();
+  }
+
+  return false; // 分析遍不修改IR
+}
+
+void SysYAliasAnalysisPass::collectMemoryAccesses(Function* F) {
+  // 收集函数中所有内存访问指令
+  for (auto& bb : F->getBasicBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      Value* ptr = nullptr;
+      
+      if (auto* loadInst = dynamic_cast<LoadInst*>(inst.get())) {
+        ptr = loadInst->getPointer();
+      } else if (auto* storeInst = dynamic_cast<StoreInst*>(inst.get())) {
+        ptr = storeInst->getPointer();
+      }
+      
+      if (ptr) {
+        // 创建内存位置信息
+        auto location = createMemoryLocation(ptr);
+        location->accessInsts.push_back(inst.get());
+        
+        // 更新读写标记
+        if (dynamic_cast<LoadInst*>(inst.get())) {
+          location->hasReads = true;
+        } else {
+          location->hasWrites = true;
+        }
+        
+        CurrentResult->addMemoryLocation(std::move(location));
+      }
+    }
+  }
+}
+
+void SysYAliasAnalysisPass::buildAliasRelations(Function *F) {
+  // 构建所有内存访问之间的别名关系
+  auto& locationMap = CurrentResult->LocationMap;
+  
+  std::vector<Value*> allPointers;
+  for (auto& pair : locationMap) {
+    allPointers.push_back(pair.first);
+  }
+  
+  // 两两比较所有指针
+  for (size_t i = 0; i < allPointers.size(); ++i) {
+    for (size_t j = i + 1; j < allPointers.size(); ++j) {
+      Value* ptr1 = allPointers[i];
+      Value* ptr2 = allPointers[j];
+      
+      MemoryLocation* loc1 = locationMap[ptr1].get();
+      MemoryLocation* loc2 = locationMap[ptr2].get();
+      
+      AliasType aliasType = analyzeAliasBetween(loc1, loc2);
+      CurrentResult->addAliasRelation(ptr1, ptr2, aliasType);
+    }
+  }
+}
+
+void SysYAliasAnalysisPass::optimizeForSysY(Function* F) {
+  // SysY特化优化
+  applySysYConstraints(F);
+  optimizeParameterAnalysis(F);
+  optimizeArrayAccessAnalysis(F);
+}
+
+std::unique_ptr<MemoryLocation> SysYAliasAnalysisPass::createMemoryLocation(Value* ptr) {
+  Value* basePtr = getBasePointer(ptr);
+  auto location = std::make_unique<MemoryLocation>(basePtr, ptr);
+  
+  // 分析内存类型和索引模式
+  analyzeMemoryType(location.get());
+  analyzeIndexPattern(location.get());
+  
+  return location;
+}
+
+Value* SysYAliasAnalysisPass::getBasePointer(Value* ptr) {
+  // 递归剥离GEP指令，找到真正的基指针
+  if (auto* gepInst = dynamic_cast<GetElementPtrInst*>(ptr)) {
+    return getBasePointer(gepInst->getBasePointer());
+  }
+  return ptr;
+}
+
+void SysYAliasAnalysisPass::analyzeMemoryType(MemoryLocation* location) {
+  Value* base = location->basePointer;
+  
+  // 检查内存类型
+  if (dynamic_cast<AllocaInst*>(base)) {
+    location->isLocalArray = true;
+  } else if (dynamic_cast<Argument*>(base)) {
+    location->isFunctionParameter = true;
+  } else if (dynamic_cast<GlobalValue*>(base)) {
+    location->isGlobalArray = true;
+  }
+}
+
+void SysYAliasAnalysisPass::analyzeIndexPattern(MemoryLocation* location) {
+  // 分析GEP指令的索引模式
+  if (auto* gepInst = dynamic_cast<GetElementPtrInst*>(location->accessPointer)) {
+    // 初始化为true，如果发现非常量索引则设为false
+    location->hasConstantIndices = true;
+    
+    // 收集所有索引
+    for (unsigned i = 0; i < gepInst->getNumIndices(); ++i) {
+      Value* index = gepInst->getIndex(i);
+      location->indices.push_back(index);
+      
+      // 检查是否为常量索引
+      if (!isConstantValue(index)) {
+        location->hasConstantIndices = false;
+      }
+    }
+    
+    // 检查是否包含循环变量
+    Function* containingFunc = nullptr;
+    if (auto* inst = dynamic_cast<Instruction*>(location->basePointer)) {
+      containingFunc = inst->getParent()->getParent();
+    } else if (auto* arg = dynamic_cast<Argument*>(location->basePointer)) {
+      containingFunc = arg->getParent();
+    }
+    
+    if (containingFunc) {
+      location->hasLoopVariableIndex = hasLoopVariableInIndices(location->indices, containingFunc);
+    }
+    
+    // 计算常量偏移
+    if (location->hasConstantIndices) {
+      location->constantOffset = calculateConstantOffset(location->indices);
+    }
+  }
+}
+
+AliasType SysYAliasAnalysisPass::analyzeAliasBetween(MemoryLocation* loc1, MemoryLocation* loc2) {
+  // 分析两个内存位置之间的别名关系
+  
+  // 1. 相同基指针的情况需要进一步分析索引
+  if (loc1->basePointer == loc2->basePointer) {
+    // 如果是同一个访问指针，那就是完全相同的内存位置
+    if (loc1->accessPointer == loc2->accessPointer) {
+      return AliasType::SELF_ALIAS;
+    }
+    
+    // 相同基指针但不同访问指针，需要比较索引
+    return compareIndices(loc1, loc2);
+  }
+  
+  // 2. 不同类型的内存位置
+  if ((loc1->isLocalArray && loc2->isLocalArray)) {
+    return compareLocalArrays(loc1, loc2);
+  }
+  
+  if ((loc1->isFunctionParameter && loc2->isFunctionParameter)) {
+    return compareParameters(loc1, loc2);
+  }
+  
+  if ((loc1->isGlobalArray || loc2->isGlobalArray)) {
+    return compareWithGlobal(loc1, loc2);
+  }
+  
+  return compareMixedTypes(loc1, loc2);
+}
+
+AliasType SysYAliasAnalysisPass::compareIndices(MemoryLocation* loc1, MemoryLocation* loc2) {
+  // 比较相同基指针下的不同索引访问
+  
+  // 如果都有常量索引，可以精确比较
+  if (loc1->hasConstantIndices && loc2->hasConstantIndices) {
+    // 比较索引数量
+    if (loc1->indices.size() != loc2->indices.size()) {
+      return AliasType::NO_ALIAS;
+    }
+    
+    // 逐个比较索引值
+    for (size_t i = 0; i < loc1->indices.size(); ++i) {
+      Value* idx1 = loc1->indices[i];
+      Value* idx2 = loc2->indices[i];
+      
+      // 都是常量，比较值
+      auto* const1 = dynamic_cast<ConstantInteger*>(idx1);
+      auto* const2 = dynamic_cast<ConstantInteger*>(idx2);
+      
+      if (const1 && const2) {
+        int val1 = std::get<int>(const1->getVal());
+        int val2 = std::get<int>(const2->getVal());
+        
+        if (val1 != val2) {
+          return AliasType::NO_ALIAS;  // 不同常量索引，确定无别名
+        }
+      } else {
+        // 不是常量，无法确定
+        return AliasType::POSSIBLE_ALIAS;
+      }
+    }
+    
+    // 所有索引都相同
+    return AliasType::SELF_ALIAS;
+  }
+  
+  // 如果有非常量索引，保守估计
+  return AliasType::POSSIBLE_ALIAS;
+}
+
+AliasType SysYAliasAnalysisPass::compareLocalArrays(MemoryLocation* loc1, MemoryLocation* loc2) {
+  // 不同局部数组不别名
+  return AliasType::NO_ALIAS;
+}
+
+AliasType SysYAliasAnalysisPass::compareParameters(MemoryLocation* loc1, MemoryLocation* loc2) {
+  // SysY特化：可配置的数组参数别名策略
+  // 
+  // SysY中数组参数的语法形式：
+  //   void func(int a[], int b[])     - 一维数组参数
+  //   void func(int a[][10], int b[]) - 多维数组参数
+  //
+  // 默认保守策略：不同数组参数可能别名（因为可能传入相同数组）
+  //   func(arr, arr);  // 传入同一个数组给两个参数
+  //
+  // 激进策略：假设不同数组参数不会传入相同数组（适用于评测环境）
+  //   在SysY评测中，这种情况很少出现
+  
+  if (useAggressiveParameterAnalysis()) {
+    // 激进策略：不同数组参数假设不别名
+    return AliasType::NO_ALIAS;
+  } else {
+    // 保守策略：不同数组参数可能别名
+    return AliasType::POSSIBLE_ALIAS;
+  }
+}
+
+AliasType SysYAliasAnalysisPass::compareWithGlobal(MemoryLocation* loc1, MemoryLocation* loc2) {
+  // 涉及全局数组的访问分析
+  // 这里处理所有涉及全局数组的情况
+  
+  // SysY特化：局部数组与全局数组不别名
+  if ((loc1->isLocalArray && loc2->isGlobalArray) || 
+      (loc1->isGlobalArray && loc2->isLocalArray)) {
+    // 局部数组在栈上，全局数组在全局区，确定不别名
+    return AliasType::NO_ALIAS;
+  }
+  
+  // SysY特化：数组参数与全局数组可能别名（保守处理）
+  if ((loc1->isFunctionParameter && loc2->isGlobalArray) || 
+      (loc1->isGlobalArray && loc2->isFunctionParameter)) {
+    // 数组参数可能指向全局数组，需要保守处理
+    return AliasType::POSSIBLE_ALIAS;
+  }
+  
+  // 其他涉及全局数组的情况，采用保守策略
+  return AliasType::POSSIBLE_ALIAS;
+}
+
+AliasType SysYAliasAnalysisPass::compareMixedTypes(MemoryLocation* loc1, MemoryLocation* loc2) {
+  // 混合类型访问的别名分析
+  // 处理不同内存类型之间的别名关系
+  
+  // SysY特化：局部数组与数组参数通常不别名
+  // 典型场景：
+  //   void func(int p[]) {       // p 是数组参数
+  //       int local[10];         // local 是局部数组
+  //       p[0] = local[0];       // 混合类型访问
+  //   }
+  // 或多维数组：
+  //   void func(int p[][10]) {   // p 是多维数组参数
+  //       int local[10];         // local 是局部数组
+  //       p[i][0] = local[0];    // 混合类型访问
+  //   } 
+  // 局部数组与数组参数：在SysY中通常不别名
+  if ((loc1->isLocalArray && loc2->isFunctionParameter) || 
+      (loc1->isFunctionParameter && loc2->isLocalArray)) {
+    // 因为局部数组是栈上分配，而数组参数是传入的外部数组
+    return AliasType::NO_ALIAS;
+  }
+
+  // 对于其他混合情况，保守估计
+  return AliasType::UNKNOWN_ALIAS;
+}
+
+void SysYAliasAnalysisPass::applySysYConstraints(Function* F) {
+  // SysY语言特定的约束和优化
+  // 1. SysY没有指针运算，简化了别名分析
+  // 2. 数组传参时保持数组语义
+  // 3. 没有动态内存分配，所有数组要么是局部的要么是参数/全局
+}
+
+void SysYAliasAnalysisPass::optimizeParameterAnalysis(Function* F) {
+  // 数组参数别名分析优化
+  // 为SysY评测环境提供可配置的优化策略
+  
+  if (!enableParameterOptimization()) {
+    return; // 保持默认的保守策略
+  }
+  
+  // 可选的参数优化：假设不同数组参数不会传入相同数组
+  // 典型的SysY函数调用：
+  //   int arr1[10], arr2[20];
+  //   func(arr1, arr2);  // 传入不同数组
+  // 而不是：
+  //   func(arr1, arr1);  // 传入相同数组给两个参数
+  // 这在SysY评测中通常是安全的假设
+  auto& locationMap = CurrentResult->LocationMap;
+  
+  for (auto it1 = locationMap.begin(); it1 != locationMap.end(); ++it1) {
+    for (auto it2 = std::next(it1); it2 != locationMap.end(); ++it2) {
+      MemoryLocation* loc1 = it1->second.get();
+      MemoryLocation* loc2 = it2->second.get();
+      
+      // 如果两个都是数组参数且基指针不同，设为NO_ALIAS
+      if (loc1->isFunctionParameter && loc2->isFunctionParameter && 
+          loc1->basePointer != loc2->basePointer) {
+        CurrentResult->addAliasRelation(it1->first, it2->first, AliasType::NO_ALIAS);
+      }
+    }
+  }
+}
+
+void SysYAliasAnalysisPass::optimizeArrayAccessAnalysis(Function* F) {
+  // 数组访问别名分析优化
+  // 基于SysY语言的特点进行简单优化
+  
+  // 优化1：同一数组的不同常量索引访问确定无别名
+  optimizeConstantIndexAccesses();
+  
+  // 优化2：识别简单的顺序访问模式
+  optimizeSequentialAccesses();
+}
+
+bool SysYAliasAnalysisPass::isConstantValue(Value* val) {
+  return dynamic_cast<ConstantInteger*>(val) != nullptr; // 简化，只检查整数常量
+}
+
+bool SysYAliasAnalysisPass::hasLoopVariableInIndices(const std::vector<Value*>& indices, Function* F) {
+  // 保守策略：所有非常量索引都视为可能的循环变量
+  // 这样可以避免复杂的循环分析依赖，保持分析的独立性
+  for (Value* index : indices) {
+    if (!isConstantValue(index)) {
+      return true; // 保守估计，确保正确性
+    }
+  }
+  return false;
+}
+
+int SysYAliasAnalysisPass::calculateConstantOffset(const std::vector<Value*>& indices) {
+  int offset = 0;
+  for (Value* index : indices) {
+    if (auto* constInt = dynamic_cast<ConstantInteger*>(index)) {
+      // ConstantInteger的getVal()返回variant，需要提取int值
+      auto val = constInt->getVal();
+      if (std::holds_alternative<int>(val)) {
+        offset += std::get<int>(val);
+      }
+    }
+  }
+  return offset;
+}
+
+void SysYAliasAnalysisPass::printStatistics() const {
+  if (CurrentResult) {
+    CurrentResult->print();
+  }
+}
+
+void SysYAliasAnalysisPass::optimizeConstantIndexAccesses() {
+  // 优化常量索引访问的别名关系
+  // 对于相同基指针的访问，如果索引都是常量且不同，则确定无别名
+  
+  auto& locationMap = CurrentResult->LocationMap;
+  std::vector<Value*> allPointers;
+  for (auto& pair : locationMap) {
+    allPointers.push_back(pair.first);
+  }
+  
+  for (size_t i = 0; i < allPointers.size(); ++i) {
+    for (size_t j = i + 1; j < allPointers.size(); ++j) {
+      Value* ptr1 = allPointers[i];
+      Value* ptr2 = allPointers[j];
+      MemoryLocation* loc1 = locationMap[ptr1].get();
+      MemoryLocation* loc2 = locationMap[ptr2].get();
+      
+      // 相同基指针且都有常量索引
+      if (loc1->basePointer == loc2->basePointer && 
+          loc1->hasConstantIndices && loc2->hasConstantIndices) {
+        
+        // 比较常量偏移
+        if (loc1->constantOffset != loc2->constantOffset) {
+          // 不同的常量偏移，确定无别名
+          CurrentResult->addAliasRelation(ptr1, ptr2, AliasType::NO_ALIAS);
+        }
+      }
+    }
+  }
+}
+
+void SysYAliasAnalysisPass::optimizeSequentialAccesses() {
+  // 识别和优化顺序访问模式
+  // 这是一个简化的实现，主要用于识别数组的顺序遍历
+  
+  // 在SysY中，大多数数组访问都是通过循环进行的
+  // 对于非常量索引的访问，我们采用保守策略，不进行过多优化
+  // 这样可以保持分析的简单性和正确性
+  
+  // 未来如果需要更精确的分析，可以在这里添加更复杂的逻辑
+}
+
+} // namespace sysy

src/midend/Pass/Analysis/CallGraphAnalysis.cpp

@@ -0,0 +1,417 @@
+#include "CallGraphAnalysis.h"
+#include "SysYIRPrinter.h"
+#include <iostream>
+#include <stack>
+#include <unordered_set>
+
+extern int DEBUG;
+
+namespace sysy {
+
+// 静态成员初始化
+void* CallGraphAnalysisPass::ID = (void*)&CallGraphAnalysisPass::ID;
+
+// ========== CallGraphAnalysisResult 实现 ==========
+
+CallGraphAnalysisResult::Statistics CallGraphAnalysisResult::getStatistics() const {
+    Statistics stats = {};
+    stats.totalFunctions = nodes.size();
+    
+    size_t totalCallEdges = 0;
+    size_t recursiveFunctions = 0;
+    size_t selfRecursiveFunctions = 0;
+    size_t totalCallers = 0;
+    size_t totalCallees = 0;
+    
+    for (const auto& pair : nodes) {
+        const auto& node = pair.second;
+        totalCallEdges += node->callees.size();
+        totalCallers += node->callers.size();
+        totalCallees += node->callees.size();
+        
+        if (node->isRecursive) recursiveFunctions++;
+        if (node->isSelfRecursive) selfRecursiveFunctions++;
+    }
+    
+    stats.totalCallEdges = totalCallEdges;
+    stats.recursiveFunctions = recursiveFunctions;
+    stats.selfRecursiveFunctions = selfRecursiveFunctions;
+    stats.stronglyConnectedComponents = sccs.size();
+    
+    // 计算最大SCC大小
+    size_t maxSCCSize = 0;
+    for (const auto& scc : sccs) {
+        maxSCCSize = std::max(maxSCCSize, scc.size());
+    }
+    stats.maxSCCSize = maxSCCSize;
+    
+    // 计算平均值
+    if (stats.totalFunctions > 0) {
+        stats.avgCallersPerFunction = static_cast<double>(totalCallers) / stats.totalFunctions;
+        stats.avgCalleesPerFunction = static_cast<double>(totalCallees) / stats.totalFunctions;
+    }
+    
+    return stats;
+}
+
+void CallGraphAnalysisResult::print() const {
+    std::cout << "---- Call Graph Analysis Results for Module ----\n";
+    
+    // 打印基本统计信息
+    auto stats = getStatistics();
+    std::cout << "  Statistics:\n";
+    std::cout << "    Total Functions: " << stats.totalFunctions << "\n";
+    std::cout << "    Total Call Edges: " << stats.totalCallEdges << "\n";
+    std::cout << "    Recursive Functions: " << stats.recursiveFunctions << "\n";
+    std::cout << "    Self-Recursive Functions: " << stats.selfRecursiveFunctions << "\n";
+    std::cout << "    Strongly Connected Components: " << stats.stronglyConnectedComponents << "\n";
+    std::cout << "    Max SCC Size: " << stats.maxSCCSize << "\n";
+    std::cout << "    Avg Callers per Function: " << stats.avgCallersPerFunction << "\n";
+    std::cout << "    Avg Callees per Function: " << stats.avgCalleesPerFunction << "\n";
+    
+    // 打印拓扑排序结果
+    std::cout << "  Topological Order (" << topologicalOrder.size() << "):\n";
+    for (size_t i = 0; i < topologicalOrder.size(); ++i) {
+        std::cout << "    " << i << ": " << topologicalOrder[i]->getName() << "\n";
+    }
+    
+    // 打印强连通分量
+    if (!sccs.empty()) {
+        std::cout << "  Strongly Connected Components:\n";
+        for (size_t i = 0; i < sccs.size(); ++i) {
+            std::cout << "    SCC " << i << " (size " << sccs[i].size() << "): ";
+            for (size_t j = 0; j < sccs[i].size(); ++j) {
+                if (j > 0) std::cout << ", ";
+                std::cout << sccs[i][j]->getName();
+            }
+            std::cout << "\n";
+        }
+    }
+    
+    // 打印每个函数的详细信息
+    std::cout << "  Function Details:\n";
+    for (const auto& pair : nodes) {
+        const auto& node = pair.second;
+        std::cout << "    Function: " << node->function->getName();
+        
+        if (node->isRecursive) {
+            std::cout << " (Recursive";
+            if (node->isSelfRecursive) std::cout << ", Self";
+            if (node->recursiveDepth >= 0) std::cout << ", Depth=" << node->recursiveDepth;
+            std::cout << ")";
+        }
+        std::cout << "\n";
+        
+        if (!node->callers.empty()) {
+            std::cout << "      Callers (" << node->callers.size() << "): ";
+            bool first = true;
+            for (Function* caller : node->callers) {
+                if (!first) std::cout << ", ";
+                std::cout << caller->getName();
+                first = false;
+            }
+            std::cout << "\n";
+        }
+        
+        if (!node->callees.empty()) {
+            std::cout << "      Callees (" << node->callees.size() << "): ";
+            bool first = true;
+            for (Function* callee : node->callees) {
+                if (!first) std::cout << ", ";
+                std::cout << callee->getName();
+                first = false;
+            }
+            std::cout << "\n";
+        }
+    }
+    
+    std::cout << "--------------------------------------------------\n";
+}
+
+void CallGraphAnalysisResult::addNode(Function* F) {
+    if (nodes.find(F) == nodes.end()) {
+        nodes[F] = std::make_unique<CallGraphNode>(F);
+    }
+}
+
+void CallGraphAnalysisResult::addCallEdge(Function* caller, Function* callee) {
+    // 确保两个函数都有对应的节点
+    addNode(caller);
+    addNode(callee);
+    
+    // 添加调用边
+    nodes[caller]->callees.insert(callee);
+    nodes[callee]->callers.insert(caller);
+    
+    // 更新统计信息
+    nodes[caller]->totalCallees = nodes[caller]->callees.size();
+    nodes[callee]->totalCallers = nodes[callee]->callers.size();
+    
+    // 检查自递归
+    if (caller == callee) {
+        nodes[caller]->isSelfRecursive = true;
+        nodes[caller]->isRecursive = true;
+    }
+}
+
+void CallGraphAnalysisResult::computeTopologicalOrder() {
+    topologicalOrder.clear();
+    std::unordered_set<Function*> visited;
+    
+    // 对每个未访问的函数进行DFS
+    for (const auto& pair : nodes) {
+        Function* F = pair.first;
+        if (visited.find(F) == visited.end()) {
+            dfsTopological(F, visited, topologicalOrder);
+        }
+    }
+    
+    // 反转结果（因为我们在后序遍历中添加）
+    std::reverse(topologicalOrder.begin(), topologicalOrder.end());
+}
+
+void CallGraphAnalysisResult::dfsTopological(Function* F, std::unordered_set<Function*>& visited, 
+                                            std::vector<Function*>& result) {
+    visited.insert(F);
+    
+    auto node = getNode(F);
+    if (node) {
+        // 先访问所有被调用的函数
+        for (Function* callee : node->callees) {
+            if (visited.find(callee) == visited.end()) {
+                dfsTopological(callee, visited, result);
+            }
+        }
+    }
+    
+    // 后序遍历：访问完所有子节点后添加当前节点
+    result.push_back(F);
+}
+
+void CallGraphAnalysisResult::computeStronglyConnectedComponents() {
+    tarjanSCC();
+    
+    // 为每个函数设置其所属的SCC
+    functionToSCC.clear();
+    for (size_t i = 0; i < sccs.size(); ++i) {
+        for (Function* F : sccs[i]) {
+            functionToSCC[F] = static_cast<int>(i);
+        }
+    }
+}
+
+void CallGraphAnalysisResult::tarjanSCC() {
+    sccs.clear();
+    
+    std::vector<int> indices(nodes.size(), -1);
+    std::vector<int> lowlinks(nodes.size(), -1);
+    std::vector<Function*> stack;
+    std::unordered_set<Function*> onStack;
+    int index = 0;
+    
+    // 为函数分配索引
+    std::map<Function*, int> functionIndex;
+    int idx = 0;
+    for (const auto& pair : nodes) {
+        functionIndex[pair.first] = idx++;
+    }
+    
+    // 对每个未访问的函数运行Tarjan算法
+    for (const auto& pair : nodes) {
+        Function* F = pair.first;
+        int fIdx = functionIndex[F];
+        if (indices[fIdx] == -1) {
+            tarjanDFS(F, index, indices, lowlinks, stack, onStack);
+        }
+    }
+}
+
+void CallGraphAnalysisResult::tarjanDFS(Function* F, int& index, std::vector<int>& indices, 
+                                      std::vector<int>& lowlinks, std::vector<Function*>& stack, 
+                                      std::unordered_set<Function*>& onStack) {
+    // 这里需要函数到索引的映射，简化实现
+    // 在实际实现中应该维护一个全局的函数索引映射
+    static std::map<Function*, int> functionIndex;
+    static int nextIndex = 0;
+    
+    if (functionIndex.find(F) == functionIndex.end()) {
+        functionIndex[F] = nextIndex++;
+    }
+    
+    int fIdx = functionIndex[F];
+    
+    // 确保向量足够大
+    if (fIdx >= static_cast<int>(indices.size())) {
+        indices.resize(fIdx + 1, -1);
+        lowlinks.resize(fIdx + 1, -1);
+    }
+    
+    indices[fIdx] = index;
+    lowlinks[fIdx] = index;
+    index++;
+    
+    stack.push_back(F);
+    onStack.insert(F);
+    
+    auto node = getNode(F);
+    if (node) {
+        for (Function* callee : node->callees) {
+            int calleeIdx = functionIndex[callee];
+            
+            // 确保向量足够大
+            if (calleeIdx >= static_cast<int>(indices.size())) {
+                indices.resize(calleeIdx + 1, -1);
+                lowlinks.resize(calleeIdx + 1, -1);
+            }
+            
+            if (indices[calleeIdx] == -1) {
+                // 递归访问
+                tarjanDFS(callee, index, indices, lowlinks, stack, onStack);
+                lowlinks[fIdx] = std::min(lowlinks[fIdx], lowlinks[calleeIdx]);
+            } else if (onStack.find(callee) != onStack.end()) {
+                // 后向边
+                lowlinks[fIdx] = std::min(lowlinks[fIdx], indices[calleeIdx]);
+            }
+        }
+    }
+    
+    // 如果F是SCC的根
+    if (lowlinks[fIdx] == indices[fIdx]) {
+        std::vector<Function*> scc;
+        Function* w;
+        do {
+            w = stack.back();
+            stack.pop_back();
+            onStack.erase(w);
+            scc.push_back(w);
+        } while (w != F);
+        
+        sccs.push_back(std::move(scc));
+    }
+}
+
+void CallGraphAnalysisResult::analyzeRecursion() {
+    // 基于SCC分析递归
+    for (const auto& scc : sccs) {
+        if (scc.size() > 1) {
+            // 多函数的SCC，标记为相互递归
+            for (Function* F : scc) {
+                auto* node = getMutableNode(F);
+                if (node) {
+                    node->isRecursive = true;
+                    node->recursiveDepth = -1; // 相互递归，深度未定义
+                }
+            }
+        } else if (scc.size() == 1) {
+            // 单函数SCC，检查是否自递归
+            Function* F = scc[0];
+            auto* node = getMutableNode(F);
+            if (node && node->callees.count(F) > 0) {
+                node->isSelfRecursive = true;
+                node->isRecursive = true;
+                node->recursiveDepth = -1; // 简化：不计算递归深度
+            }
+        }
+    }
+}
+
+// ========== CallGraphAnalysisPass 实现 ==========
+
+bool CallGraphAnalysisPass::runOnModule(Module* M, AnalysisManager& AM) {
+    if (DEBUG) {
+        std::cout << "Running Call Graph Analysis on module\n";
+    }
+    
+    // 创建分析结果
+    CurrentResult = std::make_unique<CallGraphAnalysisResult>(M);
+    
+    // 执行主要分析步骤
+    buildCallGraph(M);
+    CurrentResult->computeTopologicalOrder();
+    CurrentResult->computeStronglyConnectedComponents();
+    CurrentResult->analyzeRecursion();
+    
+    if (DEBUG) {
+        CurrentResult->print();
+    }
+    
+    return false; // 分析遍不修改IR
+}
+
+void CallGraphAnalysisPass::buildCallGraph(Module* M) {
+    // 1. 为所有函数创建节点（包括声明但未定义的函数）
+    for (auto& pair : M->getFunctions()) {
+        Function* F = pair.second.get();
+        if (!isLibraryFunction(F) && !isIntrinsicFunction(F)) {
+            CurrentResult->addNode(F);
+        }
+    }
+    
+    // 2. 扫描所有函数的调用关系
+    for (auto& pair : M->getFunctions()) {
+        Function* F = pair.second.get();
+        if (!isLibraryFunction(F) && !isIntrinsicFunction(F)) {
+            scanFunctionCalls(F);
+        }
+    }
+}
+
+void CallGraphAnalysisPass::scanFunctionCalls(Function* F) {
+    // 遍历函数中的所有基本块和指令
+    for (auto& BB : F->getBasicBlocks_NoRange()) {
+        for (auto& I : BB->getInstructions()) {
+            if (CallInst* call = dynamic_cast<CallInst*>(I.get())) {
+                processCallInstruction(call, F);
+            }
+        }
+    }
+}
+
+void CallGraphAnalysisPass::processCallInstruction(CallInst* call, Function* caller) {
+    Function* callee = call->getCallee();
+    
+    if (!callee) {
+        // 间接调用，无法静态确定目标函数
+        return;
+    }
+    
+    if (isLibraryFunction(callee) || isIntrinsicFunction(callee)) {
+        // 跳过标准库函数和内置函数
+        return;
+    }
+    
+    // 添加调用边
+    CurrentResult->addCallEdge(caller, callee);
+    
+    // 更新调用点统计
+    auto* node = CurrentResult->getMutableNode(caller);
+    if (node) {
+        node->callSiteCount++;
+    }
+}
+
+bool CallGraphAnalysisPass::isLibraryFunction(Function* F) const {
+    std::string name = F->getName();
+    
+    // SysY标准库函数
+    return name == "getint" || name == "getch" || name == "getfloat" ||
+           name == "getarray" || name == "getfarray" ||
+           name == "putint" || name == "putch" || name == "putfloat" ||
+           name == "putarray" || name == "putfarray" ||
+           name == "_sysy_starttime" || name == "_sysy_stoptime";
+}
+
+bool CallGraphAnalysisPass::isIntrinsicFunction(Function* F) const {
+    std::string name = F->getName();
+    
+    // 编译器内置函数（后续可以增加某些内置函数）
+    return name.substr(0, 5) == "llvm." || name.substr(0, 5) == "sysy.";
+}
+
+void CallGraphAnalysisPass::printStatistics() const {
+    if (CurrentResult) {
+        CurrentResult->print();
+    }
+}
+
+} // namespace sysy

src/midend/Pass/Analysis/Dom.cpp

@@ -1,21 +1,30 @@
 #include "Dom.h"
-#include <algorithm> // for std::set_intersection, std::set_difference, std::set_union
+#include <algorithm> // for std::set_intersection, std::reverse
 #include <iostream>  // for debug output
 #include <limits>    // for std::numeric_limits
 #include <queue>
+#include <functional> // for std::function
+#include <map>
+#include <vector>
+#include <set>
 
 namespace sysy {
 
-// 初始化 支配树静态 ID
+// ==============================================================
+// DominatorTreeAnalysisPass 的静态ID
+// ==============================================================
 void *DominatorTreeAnalysisPass::ID = (void *)&DominatorTreeAnalysisPass::ID;
+
 // ==============================================================
 // DominatorTree 结果类的实现
 // ==============================================================
 
+// 构造函数：初始化关联函数，但不进行计算
 DominatorTree::DominatorTree(Function *F) : AssociatedFunction(F) {
-  // 构造时可以不计算，在分析遍运行里计算并填充
+  // 构造时不需要计算，在分析遍运行里计算并填充
 }
 
+// Getter 方法 (保持不变)
 const std::set<BasicBlock *> *DominatorTree::getDominators(BasicBlock *BB) const {
   auto it = Dominators.find(BB);
   if (it != Dominators.end()) {
@@ -48,7 +57,7 @@ const std::set<BasicBlock *> *DominatorTree::getDominatorTreeChildren(BasicBlock
   return nullptr;
 }
 
-// 辅助函数：打印 BasicBlock 集合
+// 辅助函数：打印 BasicBlock 集合 (保持不变)
 void printBBSet(const std::string &prefix, const std::set<BasicBlock *> &s) {
   if (!DEBUG)
     return;
@@ -63,24 +72,52 @@ void printBBSet(const std::string &prefix, const std::set<BasicBlock *> &s) {
   std::cout << "}" << std::endl;
 }
 
+// 辅助函数：计算逆后序遍历 (RPO) - 保持不变
+std::vector<BasicBlock*> DominatorTree::computeReversePostOrder(Function* F) {
+    std::vector<BasicBlock*> postOrder;
+    std::set<BasicBlock*> visited;
+    
+    std::function<void(BasicBlock*)> dfs_rpo =
+        [&](BasicBlock* bb) {
+        visited.insert(bb);
+        for (BasicBlock* succ : bb->getSuccessors()) {
+            if (visited.find(succ) == visited.end()) {
+                dfs_rpo(succ);
+            }
+        }
+        postOrder.push_back(bb);
+    };
+
+    dfs_rpo(F->getEntryBlock());
+    std::reverse(postOrder.begin(), postOrder.end());
+    
+    if (DEBUG) {
+        std::cout << "--- Computed RPO: ";
+        for (BasicBlock* bb : postOrder) {
+            std::cout << bb->getName() << " ";
+        }
+        std::cout << "---" << std::endl;
+    }
+    return postOrder;
+}
+
+// computeDominators 方法 (保持不变，因为它它是独立于IDom算法的)
 void DominatorTree::computeDominators(Function *F) {
   if (DEBUG)
     std::cout << "--- Computing Dominators ---" << std::endl;
 
   BasicBlock *entryBlock = F->getEntryBlock();
-  std::vector<BasicBlock *> bbs_in_order; // 用于确定遍历顺序，如果需要的话
+  std::vector<BasicBlock*> bbs_rpo = computeReversePostOrder(F);
 
-  // 初始化：入口块只被自己支配，其他块被所有块支配
-  for (const auto &bb_ptr : F->getBasicBlocks()) {
-    BasicBlock *bb = bb_ptr.get();
-    bbs_in_order.push_back(bb); // 收集所有块
+  for (BasicBlock *bb : bbs_rpo) {
     if (bb == entryBlock) {
+      Dominators[bb].clear();
       Dominators[bb].insert(bb);
-      if (DEBUG)
-        std::cout << "Init Dominators[" << bb->getName() << "]: {" << bb->getName() << "}" << std::endl;
+      if (DEBUG) std::cout << "Init Dominators[" << bb->getName() << "]: {" << bb->getName() << "}" << std::endl;
     } else {
-      for (const auto &all_bb_ptr : F->getBasicBlocks()) {
-        Dominators[bb].insert(all_bb_ptr.get());
+      Dominators[bb].clear();
+      for (BasicBlock *all_bb : bbs_rpo) {
+        Dominators[bb].insert(all_bb);
       }
       if (DEBUG) {
         std::cout << "Init Dominators[" << bb->getName() << "]: ";
@@ -94,35 +131,29 @@ void DominatorTree::computeDominators(Function *F) {
   while (changed) {
     changed = false;
     iteration++;
-    if (DEBUG)
-      std::cout << "Iteration " << iteration << std::endl;
+    if (DEBUG) std::cout << "Iteration " << iteration << std::endl;
 
-    // 确保遍历顺序一致性，例如可以按照DFS或BFS顺序，或者简单的迭代器顺序
-    // 如果Function::getBasicBlocks()返回的迭代器顺序稳定，则无需bbs_in_order
-    for (const auto &bb_ptr : F->getBasicBlocks()) { // 假设这个迭代器顺序稳定
-      BasicBlock *bb = bb_ptr.get();
-      if (bb == entryBlock)
-        continue;
+    for (BasicBlock *bb : bbs_rpo) {
+      if (bb == entryBlock) continue;
 
-      // 计算所有前驱的支配者集合的交集
       std::set<BasicBlock *> newDom;
       bool firstPredProcessed = false;
 
       for (BasicBlock *pred : bb->getPredecessors()) {
-        // 确保前驱的支配者集合已经计算过
-        if (Dominators.count(pred)) {
-          if (!firstPredProcessed) {
-            newDom = Dominators[pred];
-            firstPredProcessed = true;
-          } else {
-            std::set<BasicBlock *> intersection;
-            std::set_intersection(newDom.begin(), newDom.end(), Dominators[pred].begin(), Dominators[pred].end(),
-                                  std::inserter(intersection, intersection.begin()));
-            newDom = intersection;
-          }
+        if(DEBUG){
+          std::cout << "  Processing predecessor: " << pred->getName() << std::endl;
         }
+          if (!firstPredProcessed) {
+              newDom = Dominators[pred];
+              firstPredProcessed = true;
+          } else {
+              std::set<BasicBlock *> intersection;
+              std::set_intersection(newDom.begin(), newDom.end(), Dominators[pred].begin(), Dominators[pred].end(),
+                                     std::inserter(intersection, intersection.begin()));
+              newDom = intersection;
+          }
       }
-      newDom.insert(bb); // BB 永远支配自己
+      newDom.insert(bb);
 
       if (newDom != Dominators[bb]) {
         if (DEBUG) {
@@ -140,78 +171,242 @@ void DominatorTree::computeDominators(Function *F) {
     std::cout << "--- Dominators Computation Finished ---" << std::endl;
 }
 
-void DominatorTree::computeIDoms(Function *F) {
-  if (DEBUG)
-    std::cout << "--- Computing Immediate Dominators (IDoms) ---" << std::endl;
+// ==============================================================
+// Lengauer-Tarjan 算法辅助数据结构和函数 (私有成员)
+// ==============================================================
 
-  BasicBlock *entryBlock = F->getEntryBlock();
-  IDoms[entryBlock] = nullptr; // 入口块没有即时支配者
-
-  // 遍历所有非入口块
-  for (const auto &bb_ptr : F->getBasicBlocks()) {
-    BasicBlock *bb = bb_ptr.get();
-    if (bb == entryBlock)
-      continue;
-
-    BasicBlock *currentIDom = nullptr;
-    const std::set<BasicBlock *> *domsOfBB = getDominators(bb);
-    if (!domsOfBB) {
-      if (DEBUG)
-        std::cerr << "Warning: Dominators for " << bb->getName() << " not found!" << std::endl;
-      continue;
+// DFS 遍历，填充 dfnum_map, vertex_vec, parent_map
+// 对应用户代码的 dfs 函数
+void DominatorTree::dfs_lt_helper(BasicBlock* u) {
+    dfnum_map[u] = df_counter;
+    if (df_counter >= vertex_vec.size()) { // 动态调整大小
+        vertex_vec.resize(df_counter + 1);
     }
+    vertex_vec[df_counter] = u;
+    if (DEBUG) std::cout << "  DFS: Visiting " << u->getName() << ", dfnum = " << df_counter << std::endl;
+    df_counter++;
 
-    // 遍历bb的所有严格支配者 D (即 bb 的支配者中除了 bb 自身)
-    for (BasicBlock *D_candidate : *domsOfBB) {
-      if (D_candidate == bb)
-        continue; // 跳过bb自身
-
-      bool D_candidate_is_IDom = true;
-      // 检查是否存在另一个块 X，使得 D_candidate 严格支配 X 且 X 严格支配 bb
-      // 或者更直接的，检查 D_candidate 是否被 bb 的所有其他严格支配者所支配
-      for (BasicBlock *X_other_dom : *domsOfBB) {
-        if (X_other_dom == bb || X_other_dom == D_candidate)
-          continue; // 跳过bb自身和D_candidate
-
-        // 如果 X_other_dom 严格支配 bb (它在 domsOfBB 中且不是bb自身)
-        // 并且 X_other_dom 不被 D_candidate 支配，那么 D_candidate 就不是 IDom
-        const std::set<BasicBlock *> *domsOfX_other_dom = getDominators(X_other_dom);
-        if (domsOfX_other_dom && domsOfX_other_dom->count(D_candidate)) { // X_other_dom 支配 D_candidate
-          // D_candidate 被另一个支配者 X_other_dom 支配
-          // 这说明 D_candidate 位于 X_other_dom 的“下方”，X_other_dom 更接近 bb
-          // 因此 D_candidate 不是 IDom
-          D_candidate_is_IDom = false;
-          break;
+    for (BasicBlock* v : u->getSuccessors()) {
+        if (dfnum_map.find(v) == dfnum_map.end()) { // 如果 v 未访问过
+            parent_map[v] = u;
+            if (DEBUG) std::cout << "    DFS: Setting parent[" << v->getName() << "] = " << u->getName() << std::endl;
+            dfs_lt_helper(v);
         }
-      }
-      if (D_candidate_is_IDom) {
-        currentIDom = D_candidate;
-        break; // 找到即时支配者，可以退出循环，因为它是唯一的
-      }
     }
-    IDoms[bb] = currentIDom;
-    if (DEBUG) {
-      std::cout << "  IDom[" << bb->getName() << "] = " << (currentIDom ? currentIDom->getName() : "nullptr")
-                << std::endl;
-    }
-  }
-  if (DEBUG)
-    std::cout << "--- Immediate Dominators Computation Finished ---" << std::endl;
 }
 
-/*
-for each node n in a postorder traversal of the dominator tree:
-  df[n] = empty set
-  // compute DF_local(n)
-  for each child y of n in the CFG:
-    if idom[y] != n:
-      df[n] = df[n] U {y}
-  // compute DF_up(n)
-  for each child c of n in the dominator tree:
-    for each element w in df[c]:
-      if idom[w] != n:
-        df[n] = df[n] U {w}
-*/
+// 并查集：找到集合的代表，并进行路径压缩
+// 同时更新 label，确保 label[i] 总是指向其祖先链中 sdom_map 最小的节点
+// 对应用户代码的 find 函数，也包含了 eval 的逻辑
+BasicBlock* DominatorTree::evalAndCompress_lt_helper(BasicBlock* i) {
+    if (DEBUG) std::cout << "    Eval: Processing " << i->getName() << std::endl;
+    // 如果 i 是根 (ancestor_map[i] == nullptr)
+    if (ancestor_map.find(i) == ancestor_map.end() || ancestor_map[i] == nullptr) {
+        if (DEBUG) std::cout << "      Eval: " << i->getName() << " is root, returning itself." << std::endl;
+        return i; // 根节点自身就是路径上sdom最小的，因为它没有祖先
+    }
+    
+    // 如果 i 的祖先不是根，则递归查找并进行路径压缩
+    BasicBlock* root_ancestor = evalAndCompress_lt_helper(ancestor_map[i]);
+    
+    // 路径压缩时，根据 sdom_map 比较并更新 label_map
+    // 确保 label_map[i] 存储的是 i 到 root_ancestor 路径上 sdom_map 最小的节点
+    // 注意：这里的 ancestor_map[i] 已经被递归调用压缩过一次了，所以是root_ancestor的旧路径
+    // 应该比较的是 label_map[ancestor_map[i]] 和 label_map[i]
+    if (sdom_map.count(label_map[ancestor_map[i]]) && // 确保 label_map[ancestor_map[i]] 存在 sdom
+        sdom_map.count(label_map[i]) &&                // 确保 label_map[i] 存在 sdom
+        dfnum_map[sdom_map[label_map[ancestor_map[i]]]] < dfnum_map[sdom_map[label_map[i]]]) {
+        if (DEBUG) std::cout << "      Eval: Updating label for " << i->getName() << " from " 
+                              << label_map[i]->getName() << " to " << label_map[ancestor_map[i]]->getName() << std::endl;
+        label_map[i] = label_map[ancestor_map[i]];
+    }
+    
+    ancestor_map[i] = root_ancestor; // 执行路径压缩：将 i 直接指向其所属集合的根
+    if (DEBUG) std::cout << "      Eval: Path compression for " << i->getName() << ", new ancestor = " 
+                          << (root_ancestor ? root_ancestor->getName() : "nullptr") << std::endl;
+    
+    return label_map[i]; // <-- **将这里改为返回 label_map[i]**
+}
+
+// Link 函数：将 v 加入 u 的 DFS 树子树中 (实际上是并查集操作)
+// 对应用户代码的 fa[u] = fth[u];
+void DominatorTree::link_lt_helper(BasicBlock* u_parent, BasicBlock* v_child) {
+    ancestor_map[v_child] = u_parent; // 设置并查集父节点
+    label_map[v_child] = v_child;     // 初始化 label 为自身
+    if (DEBUG) std::cout << "  Link: " << v_child->getName() << " linked to " << u_parent->getName() << std::endl;
+}
+
+// ==============================================================
+// Lengauer-Tarjan 算法实现 computeIDoms
+// ==============================================================
+void DominatorTree::computeIDoms(Function *F) {
+    if (DEBUG) std::cout << "--- Computing Immediate Dominators (IDoms) using Lengauer-Tarjan ---" << std::endl;
+
+    BasicBlock *entryBlock = F->getEntryBlock();
+
+    // 1. 初始化所有 LT 相关的数据结构
+    dfnum_map.clear();
+    vertex_vec.clear();
+    parent_map.clear();
+    sdom_map.clear();
+    idom_map.clear();
+    bucket_map.clear();
+    ancestor_map.clear();
+    label_map.clear();
+    df_counter = 0; // DFS 计数器从 0 开始
+
+    // 预分配 vertex_vec 的大小，避免频繁resize
+    vertex_vec.resize(F->getBasicBlocks().size() + 1); 
+    // 在 DFS 遍历之前，先为所有基本块初始化 sdom 和 label
+    // 这是 Lengauer-Tarjan 算法的要求，确保所有节点在 Phase 2 开始前都在 map 中
+    for (auto &bb_ptr : F->getBasicBlocks()) {
+        BasicBlock* bb = bb_ptr.get();
+        sdom_map[bb] = bb; // sdom(bb) 初始化为 bb 自身
+        label_map[bb] = bb; // label(bb) 初始化为 bb 自身 (用于 Union-Find 的路径压缩)
+    }
+    // 确保入口块也被正确初始化（如果它不在 F->getBasicBlocks() 的正常迭代中）
+    sdom_map[entryBlock] = entryBlock;
+    label_map[entryBlock] = entryBlock;
+    // Phase 1: DFS 遍历并预处理
+    // 对应用户代码的 dfs(st)
+    dfs_lt_helper(entryBlock);
+    idom_map[entryBlock] = nullptr; // 入口块没有即时支配者
+    if (DEBUG) std::cout << "  IDom[" << entryBlock->getName() << "] = nullptr" << std::endl;
+
+    if (DEBUG) std::cout << "  Sdom[" << entryBlock->getName() << "] = " << entryBlock->getName() << std::endl;
+    
+    // 初始化并查集的祖先和 label
+    for (auto const& [bb_key, dfn_val] : dfnum_map) {
+        ancestor_map[bb_key] = nullptr; // 初始为独立集合的根
+        label_map[bb_key] = bb_key;   // 初始 label 为自身
+    }
+
+    if (DEBUG) {
+        std::cout << "  --- DFS Phase Complete ---" << std::endl;
+        std::cout << "  dfnum_map:" << std::endl;
+        for (auto const& [bb, dfn] : dfnum_map) {
+            std::cout << "    " << bb->getName() << " -> " << dfn << std::endl;
+        }
+        std::cout << "  vertex_vec (by dfnum):" << std::endl;
+        for (size_t k = 0; k < df_counter; ++k) {
+            if (vertex_vec[k]) std::cout << "    [" << k << "] -> " << vertex_vec[k]->getName() << std::endl;
+        }
+        std::cout << "  parent_map:" << std::endl;
+        for (auto const& [child, parent] : parent_map) {
+            std::cout << "    " << child->getName() << " -> " << (parent ? parent->getName() : "nullptr") << std::endl;
+        }
+        std::cout << "  ------------------------" << std::endl;
+    }
+
+
+    // Phase 2: 计算半支配者 (sdom)
+    // 对应用户代码的 for (int i = dfc; i >= 2; --i) 循环的上半部分
+    // 按照 DFS 编号递减的顺序遍历所有节点 (除了 entryBlock，它的 DFS 编号是 0)
+    if (DEBUG) std::cout << "--- Phase 2: Computing Semi-Dominators (sdom) ---" << std::endl;
+    for (int i = df_counter - 1; i >= 1; --i) { // 从 DFS 编号最大的节点开始，到 1
+        BasicBlock* w = vertex_vec[i]; // 当前处理的节点
+        if (DEBUG) std::cout << "  Processing node w: " << w->getName() << " (dfnum=" << i << ")" << std::endl;
+
+
+        // 对于 w 的每个前驱 v
+        for (BasicBlock* v : w->getPredecessors()) {
+            if (DEBUG) std::cout << "    Considering predecessor v: " << v->getName() << std::endl;
+            // 如果前驱 v 未被 DFS 访问过 (即不在 dfnum_map 中)，则跳过
+            if (dfnum_map.find(v) == dfnum_map.end()) {
+                if (DEBUG) std::cout << "      Predecessor " << v->getName() << " not in DFS tree, skipping." << std::endl;
+                continue; 
+            }
+
+            // 调用 evalAndCompress 来找到 v 在其 DFS 树祖先链上具有最小 sdom 的节点
+            BasicBlock* u_with_min_sdom_on_path = evalAndCompress_lt_helper(v);
+            if (DEBUG) std::cout << "      Eval(" << v->getName() << ") returned " 
+                                  << u_with_min_sdom_on_path->getName() << std::endl;
+            if (DEBUG && sdom_map.count(u_with_min_sdom_on_path) && sdom_map.count(w)) {
+                std::cout << "      Comparing sdom: dfnum[" << sdom_map[u_with_min_sdom_on_path]->getName() << "] (" << dfnum_map[sdom_map[u_with_min_sdom_on_path]] 
+                          << ") vs dfnum[" << sdom_map[w]->getName() << "] (" << dfnum_map[sdom_map[w]] << ")" << std::endl;
+            }
+            // 比较 sdom(u) 和 sdom(w)
+            if (sdom_map.count(u_with_min_sdom_on_path) && sdom_map.count(w) &&
+                dfnum_map[sdom_map[u_with_min_sdom_on_path]] < dfnum_map[sdom_map[w]]) {
+                if (DEBUG) std::cout << "      Updating sdom[" << w->getName() << "] from " 
+                                      << sdom_map[w]->getName() << " to " 
+                                      << sdom_map[u_with_min_sdom_on_path]->getName() << std::endl;
+                sdom_map[w] = sdom_map[u_with_min_sdom_on_path]; // 更新 sdom(w)
+                if (DEBUG) std::cout << "      Sdom update applied. New sdom[" << w->getName() << "] = " << sdom_map[w]->getName() << std::endl;
+            }
+        }
+        
+        // 将 w 加入 sdom(w) 对应的桶中
+        bucket_map[sdom_map[w]].push_back(w);
+        if (DEBUG) std::cout << "    Adding " << w->getName() << " to bucket of sdom(" << w->getName() << "): " 
+                              << sdom_map[w]->getName() << std::endl;
+
+        // 将 w 的父节点加入并查集 (link 操作)
+        if (parent_map.count(w) && parent_map[w] != nullptr) {
+            link_lt_helper(parent_map[w], w);
+        }
+        
+        // Phase 3-part 1: 处理 parent[w] 的桶中所有节点，确定部分 idom
+        if (parent_map.count(w) && parent_map[w] != nullptr) {
+            BasicBlock* p = parent_map[w]; // p 是 w 的父节点
+            if (DEBUG) std::cout << "    Processing bucket for parent " << p->getName() << std::endl;
+
+            // 注意：这里需要复制桶的内容，因为原始桶在循环中会被clear
+            std::vector<BasicBlock*> nodes_in_p_bucket_copy = bucket_map[p];
+            for (BasicBlock* y : nodes_in_p_bucket_copy) {
+                if (DEBUG) std::cout << "      Processing node y from bucket: " << y->getName() << std::endl;
+                // 找到 y 在其 DFS 树祖先链上具有最小 sdom 的节点
+                BasicBlock* u = evalAndCompress_lt_helper(y);
+                if (DEBUG) std::cout << "        Eval(" << y->getName() << ") returned " << u->getName() << std::endl;
+                
+                // 确定 idom(y)
+                // if sdom(eval(y)) == sdom(parent(w)), then idom(y) = parent(w)
+                // else idom(y) = eval(y)
+                if (sdom_map.count(u) && sdom_map.count(p) &&
+                    dfnum_map[sdom_map[u]] < dfnum_map[sdom_map[p]]) {
+                    idom_map[y] = u; // 确定的 idom
+                    if (DEBUG) std::cout << "        IDom[" << y->getName() << "] set to " << u->getName() << std::endl;
+                } else {
+                    idom_map[y] = p; // p 是 y 的 idom
+                    if (DEBUG) std::cout << "        IDom[" << y->getName() << "] set to " << p->getName() << std::endl;
+                }
+            }
+            bucket_map[p].clear(); // 清空桶，防止重复处理
+            if (DEBUG) std::cout << "    Cleared bucket for parent " << p->getName() << std::endl;
+        }
+    }
+
+    // Phase 3-part 2: 最终确定 idom (处理那些 idom != sdom 的节点)
+    if (DEBUG) std::cout << "--- Phase 3: Finalizing Immediate Dominators (idom) ---" << std::endl;
+    for (int i = 1; i < df_counter; ++i) { // 从 DFS 编号最小的节点 (除了 entryBlock) 开始
+        BasicBlock* w = vertex_vec[i];
+        if (DEBUG) std::cout << "  Finalizing node w: " << w->getName() << std::endl;
+        if (idom_map.count(w) && sdom_map.count(w) && idom_map[w] != sdom_map[w]) {
+            // idom[w] 的 idom 是其真正的 idom
+            if (DEBUG) std::cout << "    idom[" << w->getName() << "] (" << idom_map[w]->getName() 
+                                  << ") != sdom[" << w->getName() << "] (" << sdom_map[w]->getName() << ")" << std::endl;
+            if (idom_map.count(idom_map[w])) {
+                idom_map[w] = idom_map[idom_map[w]];
+                if (DEBUG) std::cout << "    Updating idom[" << w->getName() << "] to idom(idom(w)): " 
+                                      << idom_map[w]->getName() << std::endl;
+            } else {
+                 if (DEBUG) std::cout << "    Warning: idom(idom(" << w->getName() << ")) not found, leaving idom[" << w->getName() << "] as is." << std::endl;
+            }
+        }
+        if (DEBUG) {
+            std::cout << "  Final IDom[" << w->getName() << "] = " << (idom_map[w] ? idom_map[w]->getName() : "nullptr") << std::endl;
+        }
+    }
+
+    // 将计算结果从 idom_map 存储到 DominatorTree 的成员变量 IDoms 中
+    IDoms = idom_map; 
+
+    if (DEBUG) std::cout << "--- Immediate Dominators Computation Finished ---" << std::endl;
+}
+
+// ==============================================================
+// computeDominanceFrontiers 和 computeDominatorTreeChildren (保持不变)
+// ==============================================================
 
 void DominatorTree::computeDominanceFrontiers(Function *F) {
   if (DEBUG)
@@ -221,21 +416,17 @@ void DominatorTree::computeDominanceFrontiers(Function *F) {
     BasicBlock *X = bb_ptr_X.get();
     DominanceFrontiers[X].clear();
 
-    // 遍历所有可能的 Z (X支配Z，或者Z就是X)
     for (const auto &bb_ptr_Z : F->getBasicBlocks()) {
       BasicBlock *Z = bb_ptr_Z.get();
       const std::set<BasicBlock *> *domsOfZ = getDominators(Z);
 
-      // 如果 X 不支配 Z，则 Z 与 DF(X) 无关
-      if (!domsOfZ || domsOfZ->find(X) == domsOfZ->end()) {
+      if (!domsOfZ || domsOfZ->find(X) == domsOfZ->end()) { // Z 不被 X 支配
         continue;
       }
 
-      // 遍历 Z 的所有后继 Y
       for (BasicBlock *Y : Z->getSuccessors()) {
-        // 如果 Y 不被 X 严格支配，则 Y 在 DF(X) 中
-        // Y 不被 X 严格支配意味着 (Y不被X支配) 或 (Y就是X)
         const std::set<BasicBlock *> *domsOfY = getDominators(Y);
+        // 如果 Y == X，或者 Y 不被 X 严格支配 (即 Y 不被 X 支配)
         if (Y == X || (domsOfY && domsOfY->find(X) == domsOfY->end())) {
           DominanceFrontiers[X].insert(Y);
         }
@@ -274,23 +465,21 @@ void DominatorTree::computeDominatorTreeChildren(Function *F) {
 }
 
 // ==============================================================
-// DominatorTreeAnalysisPass 的实现
+// DominatorTreeAnalysisPass 的实现 (保持不变)
 // ==============================================================
 
 bool DominatorTreeAnalysisPass::runOnFunction(Function *F, AnalysisManager &AM) {
   // 每次运行时清空旧数据，确保重新计算
   CurrentDominatorTree = std::make_unique<DominatorTree>(F);
-  // 不需要手动清空map，unique_ptr会创建新的DominatorTree对象，其map是空的
 
   CurrentDominatorTree->computeDominators(F);
-  CurrentDominatorTree->computeIDoms(F); // 修正后的IDoms算法
+  CurrentDominatorTree->computeIDoms(F); // 修正后的LT算法
   CurrentDominatorTree->computeDominanceFrontiers(F);
   CurrentDominatorTree->computeDominatorTreeChildren(F);
-  return false; // 分析遍通常返回 false，表示不修改 IR
+  return false;
 }
 
 std::unique_ptr<AnalysisResultBase> DominatorTreeAnalysisPass::getResult() {
-  // 返回计算好的 DominatorTree 实例，所有权转移给 AnalysisManager
   return std::move(CurrentDominatorTree);
 }
 

src/midend/Pass/Analysis/Loop.cpp

@@ -0,0 +1,415 @@
+#include "Dom.h" // 确保包含 DominatorTreeAnalysisPass 的定义
+#include "Loop.h" //
+#include "AliasAnalysis.h" // 添加别名分析依赖
+#include "SideEffectAnalysis.h" // 添加副作用分析依赖
+#include <iostream>
+#include <queue> // 用于 BFS 遍历设置循环层级
+
+// 调试模式开关
+#ifndef DEBUG
+#define DEBUG 0
+#endif
+
+namespace sysy {
+
+// 定义 Pass 的唯一 ID
+void *LoopAnalysisPass::ID = (void *)&LoopAnalysisPass::ID;
+
+// 定义 Loop 类的静态变量
+int Loop::NextLoopID = 0;
+// **实现 LoopAnalysisResult::print() 方法**
+
+
+void LoopAnalysisResult::printBBSet(const std::string &prefix, const std::set<BasicBlock *> &s) const{
+  if (!DEBUG) return;
+  std::cout << prefix << "{";
+  bool first = true;
+  for (const auto &bb : s) {
+    if (!first) std::cout << ", ";
+    std::cout << bb->getName();
+    first = false;
+  }
+  std::cout << "}";
+}
+
+// **辅助函数：打印 Loop 指针向量**
+void LoopAnalysisResult::printLoopVector(const std::string &prefix, const std::vector<Loop *> &loops) const {
+  if (!DEBUG) return;
+  std::cout << prefix << "[";
+  bool first = true;
+  for (const auto &loop : loops) {
+    if (!first) std::cout << ", ";
+    std::cout << loop->getName(); // 假设 Loop::getName() 存在
+    first = false;
+  }
+  std::cout << "]";
+}
+
+void LoopAnalysisResult::print() const {
+  if (!DEBUG) return; // 只有在 DEBUG 模式下才打印
+
+  std::cout << "\n--- Loop Analysis Results for Function: " << AssociatedFunction->getName() << " ---" << std::endl;
+
+  if (AllLoops.empty()) {
+    std::cout << "  No loops found." << std::endl;
+    return;
+  }
+
+  std::cout << "Total Loops Found: " << AllLoops.size() << std::endl;
+
+  // 1. 按层级分组循环
+  std::map<int, std::vector<Loop*>> loopsByLevel;
+  int maxLevel = 0;
+  for (const auto& loop_ptr : AllLoops) {
+      if (loop_ptr->getLoopLevel() != -1) { // 确保层级已计算
+          loopsByLevel[loop_ptr->getLoopLevel()].push_back(loop_ptr.get());
+          if (loop_ptr->getLoopLevel() > maxLevel) {
+              maxLevel = loop_ptr->getLoopLevel();
+          }
+      }
+  }
+
+  // 2. 打印循环层次结构
+  std::cout << "\n--- Loop Hierarchy ---" << std::endl;
+  for (int level = 0; level <= maxLevel; ++level) {
+      if (loopsByLevel.count(level)) {
+          std::cout << "Level " << level << " Loops:" << std::endl;
+          for (Loop* loop : loopsByLevel[level]) {
+              std::string indent(level * 2, ' '); // 根据层级缩进
+              std::cout << indent << "- Loop Header: " << loop->getName() << std::endl;
+              std::cout << indent << "  Blocks: ";
+              printBBSet("", loop->getBlocks());
+              std::cout << std::endl;
+
+              std::cout << indent << "  Exit Blocks: ";
+              printBBSet("", loop->getExitBlocks());
+              std::cout << std::endl;
+
+              std::cout << indent << "  Pre-Header: " << (loop->getPreHeader() ? loop->getPreHeader()->getName() : "None") << std::endl;
+              std::cout << indent << "  Parent Loop: " << (loop->getParentLoop() ? loop->getParentLoop()->getName() : "None (Outermost)") << std::endl;
+              std::cout << indent << "  Nested Loops: ";
+              printLoopVector("", loop->getNestedLoops());
+              std::cout << std::endl;
+          }
+      }
+  }
+
+  // 3. 打印最外层/最内层循环摘要
+  std::cout << "\n--- Loop Summary ---" << std::endl;
+  std::cout << "Outermost Loops: ";
+  printLoopVector("", getOutermostLoops());
+  std::cout << std::endl;
+
+  std::cout << "Innermost Loops: ";
+  printLoopVector("", getInnermostLoops());
+  std::cout << std::endl;
+
+  std::cout << "-----------------------------------------------" << std::endl;
+}
+
+bool LoopAnalysisPass::runOnFunction(Function *F, AnalysisManager &AM) {
+  if (F->getBasicBlocks().empty()) {
+    CurrentResult = std::make_unique<LoopAnalysisResult>(F);
+    return false; // 空函数，没有循环
+  }
+
+  if (DEBUG)
+    std::cout << "Running LoopAnalysisPass on function: " << F->getName() << std::endl;
+
+  // 获取支配树分析结果
+  // 这是循环分析的关键依赖
+  DominatorTree *DT = AM.getAnalysisResult<DominatorTree, DominatorTreeAnalysisPass>(F);
+  if (!DT) {
+    // 无法获取支配树，无法进行循环分析
+    std::cerr << "Error: DominatorTreeAnalysisResult not available for function " << F->getName() << std::endl;
+    CurrentResult = std::make_unique<LoopAnalysisResult>(F);
+    return false;
+  }
+
+  // 获取别名分析结果 - 用于循环内存访问分析
+  AliasAnalysisResult *aliasAnalysis = AM.getAnalysisResult<AliasAnalysisResult, SysYAliasAnalysisPass>(F);
+  if (DEBUG && aliasAnalysis) {
+    std::cout << "Loop Analysis: Using alias analysis results for enhanced memory pattern detection" << std::endl;
+  }
+
+  // 获取副作用分析结果 - 用于循环纯度分析
+  SideEffectAnalysisResult *sideEffectAnalysis = AM.getAnalysisResult<SideEffectAnalysisResult, SysYSideEffectAnalysisPass>();
+  if (DEBUG && sideEffectAnalysis) {
+    std::cout << "Loop Analysis: Using side effect analysis results for loop purity detection" << std::endl;
+  }
+
+  CurrentResult = std::make_unique<LoopAnalysisResult>(F);
+  bool changed = false; // 循环分析本身不修改IR，所以通常返回false
+
+  // 步骤 1: 识别回边和对应的自然循环
+  // 回边 (N -> D) 定义：D 支配 N
+  std::vector<std::pair<BasicBlock *, BasicBlock *>> backEdges;
+  for (auto &BB : F->getBasicBlocks()) {
+    auto Block = BB.get();
+    for (BasicBlock *Succ : Block->getSuccessors()) {
+      if (DT->getDominators(Block) && DT->getDominators(Block)->count(Succ)) {
+        // Succ 支配 Block，所以 (Block -> Succ) 是一条回边
+        backEdges.push_back({Block, Succ});
+        if (DEBUG)
+          std::cout << "Found back edge: " << Block->getName() << " -> " << Succ->getName() << std::endl;
+      }
+    }
+  }
+  
+  if (DEBUG)
+    std::cout << "Total back edges found: " << backEdges.size() << std::endl;
+
+  // 步骤 2: 为每条回边构建自然循环
+  std::map<BasicBlock*, std::unique_ptr<Loop>> loopMap; // 按循环头分组
+  
+  for (auto &edge : backEdges) {
+    BasicBlock *N = edge.first;  // 回边的尾部
+    BasicBlock *D = edge.second; // 回边的头部 (循环头)
+
+    // 检查是否已经为此循环头创建了循环
+    if (loopMap.find(D) == loopMap.end()) {
+      // 创建新的 Loop 对象
+      loopMap[D] = std::make_unique<Loop>(D);
+    }
+    
+    Loop* currentLoop = loopMap[D].get();
+
+    // 收集此回边对应的循环体块：从 N 逆向遍历到 D
+    std::set<BasicBlock *> loopBlocks; // 临时存储循环块
+    std::queue<BasicBlock *> q;
+
+    // 循环头总是循环体的一部分
+    loopBlocks.insert(D);
+    
+    // 如果回边的尾部不是循环头本身，则将其加入队列进行遍历
+    if (N != D) {
+      q.push(N);
+      loopBlocks.insert(N);
+    }
+
+    while (!q.empty()) {
+      BasicBlock *current = q.front();
+      q.pop();
+
+      for (BasicBlock *pred : current->getPredecessors()) {
+        // 如果前驱还没有被访问过，则将其加入循环体并继续遍历
+        if (loopBlocks.find(pred) == loopBlocks.end()) {
+          loopBlocks.insert(pred);
+          q.push(pred);
+        }
+      }
+    }
+
+    // 将收集到的块添加到 Loop 对象中（合并所有回边的结果）
+    for (BasicBlock *loopBB : loopBlocks) {
+      currentLoop->addBlock(loopBB);
+    }
+  }
+
+  // 处理每个合并后的循环
+  for (auto &[header, currentLoop] : loopMap) {
+    const auto &loopBlocks = currentLoop->getBlocks();
+
+    // 步骤 3: 识别循环出口块 (Exit Blocks)
+    for (BasicBlock *loopBB : loopBlocks) {
+      for (BasicBlock *succ : loopBB->getSuccessors()) {
+        if (loopBlocks.find(succ) == loopBlocks.end()) {
+          // 如果后继不在循环体内，则 loopBB 是一个出口块
+          currentLoop->addExitBlock(loopBB);
+        }
+      }
+    }
+
+    // 步骤 4: 识别循环前置块 (Pre-Header)
+    BasicBlock *candidatePreHeader = nullptr;
+    int externalPredecessorCount = 0;
+    for (BasicBlock *predOfHeader : header->getPredecessors()) {
+      // 使用 currentLoop->contains() 来检查前驱是否在循环体内
+      if (!currentLoop->contains(predOfHeader)) {
+        // 如果前驱不在循环体内，则是一个外部前驱
+        externalPredecessorCount++;
+        candidatePreHeader = predOfHeader;
+      }
+    }
+
+    if (externalPredecessorCount == 1) {
+      currentLoop->setPreHeader(candidatePreHeader);
+    }
+    CurrentResult->addLoop(std::move(currentLoop));
+  }
+
+  // 步骤 5: 处理嵌套循环 (确定父子关系和层级)
+  const auto &allLoops = CurrentResult->getAllLoops();
+
+  // 1. 首先，清除所有循环已设置的父子关系和嵌套子循环列表，确保重新计算
+  for (const auto &loop_ptr : allLoops) {
+      loop_ptr->setParentLoop(nullptr); // 清除父指针
+      loop_ptr->clearNestedLoops();     // 清除子循环列表
+      loop_ptr->setLoopLevel(-1);       // 重置循环层级
+  }
+
+  // 2. 遍历所有循环，为每个循环找到其直接父循环并建立关系
+  for (const auto &innerLoop_ptr : allLoops) {
+    Loop *innerLoop = innerLoop_ptr.get();
+    Loop *immediateParent = nullptr; // 用于存储当前 innerLoop 的最近父循环
+
+    for (const auto &outerLoop_ptr : allLoops) {
+      Loop *outerLoop = outerLoop_ptr.get();
+
+      // 一个循环不能是它自己的父循环
+      if (outerLoop == innerLoop) {
+        continue;
+      }
+
+      // 检查 outerLoop 是否包含 innerLoop 的所有条件：
+      // Condition 1: outerLoop 的头支配 innerLoop 的头
+      if (!(DT->getDominators(innerLoop->getHeader()) &&
+            DT->getDominators(innerLoop->getHeader())->count(outerLoop->getHeader()))) {
+        continue; // outerLoop 不支配 innerLoop 的头，因此不是一个外层循环
+      }
+
+      // Condition 2: innerLoop 的所有基本块都在 outerLoop 的基本块集合中
+      bool allInnerBlocksInOuter = true;
+      for (BasicBlock *innerBB : innerLoop->getBlocks()) {
+        if (!outerLoop->contains(innerBB)) { //
+          allInnerBlocksInOuter = false;
+          break;
+        }
+      }
+      if (!allInnerBlocksInOuter) {
+        continue; // outerLoop 不包含 innerLoop 的所有块
+      }
+
+      // 到此为止，outerLoop 已经被确认为 innerLoop 的一个“候选父循环”（即它包含了 innerLoop）
+
+      if (immediateParent == nullptr) {
+        // 这是找到的第一个候选父循环
+        immediateParent = outerLoop;
+      } else {
+        // 已经有了一个 immediateParent，需要判断哪个是更“紧密”的父循环
+        // 更紧密的父循环是那个包含另一个候选父循环的。
+        // 如果当前的 immediateParent 包含了 outerLoop 的头，那么 outerLoop 是更深的循环（更接近 innerLoop）
+        if (immediateParent->contains(outerLoop->getHeader())) { //
+          immediateParent = outerLoop; // outerLoop 是更紧密的父循环
+        }
+        // 否则（outerLoop 包含了 immediateParent 的头），说明 immediateParent 更紧密，保持不变
+        // 或者它们互不包含（不应该发生，因为它们都包含了 innerLoop），也保持 immediateParent
+      }
+    }
+
+    // 设置 innerLoop 的直接父循环，并添加到父循环的嵌套列表中
+    if (immediateParent) {
+      innerLoop->setParentLoop(immediateParent);
+      immediateParent->addNestedLoop(innerLoop); 
+    }
+  }
+
+  // 3. 计算循环层级 (Level)
+  std::queue<Loop *> q_level;
+
+  // 查找所有最外层循环（没有父循环的），设置其层级为0，并加入队列
+  for (const auto &loop_ptr : allLoops) {
+    if (loop_ptr->isOutermost()) {
+      loop_ptr->setLoopLevel(0);
+      q_level.push(loop_ptr.get());
+    }
+  }
+
+  // 使用 BFS 遍历循环树，计算所有嵌套循环的层级
+  while (!q_level.empty()) {
+    Loop *current = q_level.front();
+    q_level.pop();
+
+    for (Loop *nestedLoop : current->getNestedLoops()) {
+      nestedLoop->setLoopLevel(current->getLoopLevel() + 1);
+      q_level.push(nestedLoop);
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "Loop Analysis completed for function: " << F->getName() << std::endl;
+    std::cout << "Total loops found: " << CurrentResult->getLoopCount() << std::endl;
+    std::cout << "Max loop depth: " << CurrentResult->getMaxLoopDepth() << std::endl;
+    std::cout << "Innermost loops: " << CurrentResult->getInnermostLoops().size() << std::endl;
+    std::cout << "Outermost loops: " << CurrentResult->getOutermostLoops().size() << std::endl;
+    
+    // 打印各深度的循环分布
+    for (int depth = 1; depth <= CurrentResult->getMaxLoopDepth(); ++depth) {
+      int count = CurrentResult->getLoopCountAtDepth(depth);
+      if (count > 0) {
+        std::cout << "Loops at depth " << depth << ": " << count << std::endl;
+      }
+    }
+    
+    // 输出缓存统计
+    auto cacheStats = CurrentResult->getCacheStats();
+    std::cout << "Cache statistics - Total cached queries: " << cacheStats.totalCachedQueries << std::endl;
+  }
+
+  return changed;
+}
+
+// ========== Loop 类的新增方法实现 ==========
+
+bool Loop::mayHaveSideEffects(SideEffectAnalysisResult* sideEffectAnalysis) const {
+  if (!sideEffectAnalysis) return true; // 保守假设
+  
+  for (BasicBlock* bb : LoopBlocks) {
+    for (auto& inst : bb->getInstructions()) {
+      if (sideEffectAnalysis->hasSideEffect(inst.get())) {
+        return true;
+      }
+    }
+  }
+  return false;
+}
+
+bool Loop::accessesGlobalMemory(AliasAnalysisResult* aliasAnalysis) const {
+  if (!aliasAnalysis) return true; // 保守假设
+  
+  for (BasicBlock* bb : LoopBlocks) {
+    for (auto& inst : bb->getInstructions()) {
+      if (auto* loadInst = dynamic_cast<LoadInst*>(inst.get())) {
+        if (!aliasAnalysis->isLocalArray(loadInst->getPointer())) {
+          return true;
+        }
+      } else if (auto* storeInst = dynamic_cast<StoreInst*>(inst.get())) {
+        if (!aliasAnalysis->isLocalArray(storeInst->getPointer())) {
+          return true;
+        }
+      }
+    }
+  }
+  return false;
+}
+
+bool Loop::hasMemoryAliasConflicts(AliasAnalysisResult* aliasAnalysis) const {
+  if (!aliasAnalysis) return true; // 保守假设
+  
+  std::vector<Value*> memoryAccesses;
+  
+  // 收集所有内存访问
+  for (BasicBlock* bb : LoopBlocks) {
+    for (auto& inst : bb->getInstructions()) {
+      if (auto* loadInst = dynamic_cast<LoadInst*>(inst.get())) {
+        memoryAccesses.push_back(loadInst->getPointer());
+      } else if (auto* storeInst = dynamic_cast<StoreInst*>(inst.get())) {
+        memoryAccesses.push_back(storeInst->getPointer());
+      }
+    }
+  }
+  
+  // 检查两两之间是否有别名
+  for (size_t i = 0; i < memoryAccesses.size(); ++i) {
+    for (size_t j = i + 1; j < memoryAccesses.size(); ++j) {
+      auto aliasType = aliasAnalysis->queryAlias(memoryAccesses[i], memoryAccesses[j]);
+      if (aliasType == AliasType::SELF_ALIAS || aliasType == AliasType::POSSIBLE_ALIAS) {
+        return true;
+      }
+    }
+  }
+  
+  return false;
+}
+
+} // namespace sysy

src/midend/Pass/Analysis/LoopVectorization.cpp

@@ -0,0 +1,803 @@
+#include "LoopVectorization.h"
+#include "Dom.h"
+#include "Loop.h"
+#include "Liveness.h"
+#include "AliasAnalysis.h" 
+#include "SideEffectAnalysis.h"
+#include <iostream>
+#include <algorithm>
+#include <cmath>
+#include <set>
+
+extern int DEBUG;
+
+namespace sysy {
+
+// 定义 Pass 的唯一 ID
+void *LoopVectorizationPass::ID = (void *)&LoopVectorizationPass::ID;
+
+std::vector<int> DependenceVector::getDirectionVector() const {
+  std::vector<int> direction;
+  direction.reserve(distances.size());
+  
+  for (int dist : distances) {
+    if (dist > 0) direction.push_back(1);       // 前向依赖
+    else if (dist < 0) direction.push_back(-1); // 后向依赖  
+    else direction.push_back(0);                // 无依赖
+  }
+  
+  return direction;
+}
+
+bool DependenceVector::isVectorizationSafe() const {
+  if (!isKnown) return false; // 未知依赖，不安全
+  
+  // 对于向量化，我们主要关心最内层循环的依赖
+  if (distances.empty()) return true;
+  
+  int innermostDistance = distances.back(); // 最内层循环的距离
+  
+  // 前向依赖 (距离 > 0) 通常是安全的，可以通过调整向量化顺序处理
+  // 后向依赖 (距离 < 0) 通常不安全，会阻止向量化
+  // 距离 = 0 表示同一迭代内的依赖，通常安全
+  
+  return innermostDistance >= 0;
+}
+
+size_t LoopVectorizationResult::getVectorizableLoopCount() const {
+  size_t count = 0;
+  for (const auto& [loop, analysis] : VectorizationMap) {
+    if (analysis.isVectorizable) count++;
+  }
+  return count;
+}
+
+size_t LoopVectorizationResult::getParallelizableLoopCount() const {
+  size_t count = 0;
+  for (const auto& [loop, analysis] : ParallelizationMap) {
+    if (analysis.isParallelizable) count++;
+  }
+  return count;
+}
+
+std::vector<Loop*> LoopVectorizationResult::getVectorizationCandidates() const {
+  std::vector<Loop*> candidates;
+  for (const auto& [loop, analysis] : VectorizationMap) {
+    if (analysis.isVectorizable) {
+      candidates.push_back(loop);
+    }
+  }
+  
+  // 按建议的向量宽度排序，优先处理收益更大的循环
+  std::sort(candidates.begin(), candidates.end(), 
+    [this](Loop* a, Loop* b) {
+      const auto& analysisA = VectorizationMap.at(a);
+      const auto& analysisB = VectorizationMap.at(b);
+      return analysisA.suggestedVectorWidth > analysisB.suggestedVectorWidth;
+    });
+  
+  return candidates;
+}
+
+std::vector<Loop*> LoopVectorizationResult::getParallelizationCandidates() const {
+  std::vector<Loop*> candidates;
+  for (const auto& [loop, analysis] : ParallelizationMap) {
+    if (analysis.isParallelizable) {
+      candidates.push_back(loop);
+    }
+  }
+  
+  // 按建议的线程数排序
+  std::sort(candidates.begin(), candidates.end(),
+    [this](Loop* a, Loop* b) {
+      const auto& analysisA = ParallelizationMap.at(a);
+      const auto& analysisB = ParallelizationMap.at(b);
+      return analysisA.suggestedThreadCount > analysisB.suggestedThreadCount;
+    });
+  
+  return candidates;
+}
+
+void LoopVectorizationResult::print() const {
+  if (!DEBUG) return;
+
+  std::cout << "\n--- Loop Vectorization/Parallelization Analysis Results for Function: " 
+            << AssociatedFunction->getName() << " ---" << std::endl;
+
+  if (VectorizationMap.empty() && ParallelizationMap.empty()) {
+    std::cout << "  No vectorization/parallelization analysis results." << std::endl;
+    return;
+  }
+
+  // 统计信息
+  std::cout << "\n=== Summary ===" << std::endl;
+  std::cout << "Total Loops Analyzed: " << VectorizationMap.size() << std::endl;
+  std::cout << "Vectorizable Loops: " << getVectorizableLoopCount() << std::endl;
+  std::cout << "Parallelizable Loops: " << getParallelizableLoopCount() << std::endl;
+
+  // 详细分析结果
+  for (const auto& [loop, vecAnalysis] : VectorizationMap) {
+    std::cout << "\n--- Loop: " << loop->getName() << " ---" << std::endl;
+    
+    // 向量化分析 (暂时搁置)
+    std::cout << "  Vectorization: " << (vecAnalysis.isVectorizable ? "YES" : "NO") << std::endl;
+    if (!vecAnalysis.preventingFactors.empty()) {
+      std::cout << "    Preventing Factors: ";
+      for (const auto& factor : vecAnalysis.preventingFactors) {
+        std::cout << factor << " ";
+      }
+      std::cout << std::endl;
+    }
+    
+    // 并行化分析
+    auto parallelIt = ParallelizationMap.find(loop);
+    if (parallelIt != ParallelizationMap.end()) {
+      const auto& parAnalysis = parallelIt->second;
+      std::cout << "  Parallelization: " << (parAnalysis.isParallelizable ? "YES" : "NO") << std::endl;
+      if (parAnalysis.isParallelizable) {
+        std::cout << "    Suggested Thread Count: " << parAnalysis.suggestedThreadCount << std::endl;
+        if (parAnalysis.requiresReduction) {
+          std::cout << "    Requires Reduction: Yes" << std::endl;
+        }
+        if (parAnalysis.requiresBarrier) {
+          std::cout << "    Requires Barrier: Yes" << std::endl;
+        }
+      } else if (!parAnalysis.preventingFactors.empty()) {
+        std::cout << "    Preventing Factors: ";
+        for (const auto& factor : parAnalysis.preventingFactors) {
+          std::cout << factor << " ";
+        }
+        std::cout << std::endl;
+      }
+    }
+    
+    // 依赖关系
+    auto depIt = DependenceMap.find(loop);
+    if (depIt != DependenceMap.end()) {
+      const auto& dependences = depIt->second;
+      std::cout << "  Dependences: " << dependences.size() << " found" << std::endl;
+      for (const auto& dep : dependences) {
+        if (dep.dependenceVector.isKnown) {
+          std::cout << "    " << dep.source->getName() << " -> " << dep.sink->getName();
+          std::cout << " [";
+          for (size_t i = 0; i < dep.dependenceVector.distances.size(); ++i) {
+            if (i > 0) std::cout << ",";
+            std::cout << dep.dependenceVector.distances[i];
+          }
+          std::cout << "]" << std::endl;
+        }
+      }
+    }
+  }
+  
+  std::cout << "-----------------------------------------------" << std::endl;
+}
+
+bool LoopVectorizationPass::runOnFunction(Function *F, AnalysisManager &AM) {
+  if (F->getBasicBlocks().empty()) {
+    CurrentResult = std::make_unique<LoopVectorizationResult>(F);
+    return false;
+  }
+
+  if (DEBUG) {
+    std::cout << "Running LoopVectorizationPass on function: " << F->getName() << std::endl;
+  }
+
+  // 获取循环分析结果
+  auto* loopAnalysisResult = AM.getAnalysisResult<LoopAnalysisResult, LoopAnalysisPass>(F);
+  if (!loopAnalysisResult || !loopAnalysisResult->hasLoops()) {
+    CurrentResult = std::make_unique<LoopVectorizationResult>(F);
+    return false;
+  }
+
+  // 获取循环特征分析结果
+  auto* loopCharacteristics = AM.getAnalysisResult<LoopCharacteristicsResult, LoopCharacteristicsPass>(F);
+  if (!loopCharacteristics) {
+    if (DEBUG) {
+      std::cout << "Warning: LoopCharacteristics analysis not available" << std::endl;
+    }
+  }
+
+  // 获取别名分析结果
+  auto* aliasAnalysis = AM.getAnalysisResult<AliasAnalysisResult, SysYAliasAnalysisPass>(F);
+  
+  // 获取副作用分析结果
+  auto* sideEffectAnalysis = AM.getAnalysisResult<SideEffectAnalysisResult, SysYSideEffectAnalysisPass>();
+
+  CurrentResult = std::make_unique<LoopVectorizationResult>(F);
+
+  // 分析每个循环的向量化/并行化可行性
+  for (const auto& loop_ptr : loopAnalysisResult->getAllLoops()) {
+    Loop* loop = loop_ptr.get();
+    
+    // 获取该循环的特征信息
+    LoopCharacteristics* characteristics = nullptr;
+    if (loopCharacteristics) {
+      characteristics = const_cast<LoopCharacteristics*>(loopCharacteristics->getCharacteristics(loop));
+    }
+    
+    analyzeLoop(loop, characteristics, aliasAnalysis, sideEffectAnalysis);
+  }
+
+  if (DEBUG) {
+    std::cout << "LoopVectorizationPass completed. Found " 
+              << CurrentResult->getVectorizableLoopCount() << " vectorizable loops, "
+              << CurrentResult->getParallelizableLoopCount() << " parallelizable loops" << std::endl;
+  }
+
+  return false; // 分析遍不修改IR
+}
+
+void LoopVectorizationPass::analyzeLoop(Loop* loop, LoopCharacteristics* characteristics, 
+                                       AliasAnalysisResult* aliasAnalysis, SideEffectAnalysisResult* sideEffectAnalysis) {
+  if (DEBUG) {
+    std::cout << "  Analyzing advanced features for loop: " << loop->getName() << std::endl;
+  }
+
+  // 1. 计算精确依赖向量
+  auto dependences = computeDependenceVectors(loop, aliasAnalysis);
+  CurrentResult->addDependenceAnalysis(loop, dependences);
+
+  // 2. 分析向量化可行性 (暂时搁置，总是返回不可向量化)
+  auto vecAnalysis = analyzeVectorizability(loop, dependences, characteristics);
+  CurrentResult->addVectorizationAnalysis(loop, vecAnalysis);
+
+  // 3. 分析并行化可行性 
+  auto parAnalysis = analyzeParallelizability(loop, dependences, characteristics);
+  CurrentResult->addParallelizationAnalysis(loop, parAnalysis);
+}
+
+// ========== 依赖向量分析实现 ==========
+
+std::vector<PreciseDependence> LoopVectorizationPass::computeDependenceVectors(Loop* loop, 
+                                                                              AliasAnalysisResult* aliasAnalysis) {
+  std::vector<PreciseDependence> dependences;
+  std::vector<Instruction*> memoryInsts;
+  
+  // 收集所有内存操作指令
+  for (BasicBlock* bb : loop->getBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      if (dynamic_cast<LoadInst*>(inst.get()) || dynamic_cast<StoreInst*>(inst.get())) {
+        memoryInsts.push_back(inst.get());
+      }
+    }
+  }
+  
+  // 分析每对内存操作之间的依赖关系
+  for (size_t i = 0; i < memoryInsts.size(); ++i) {
+    for (size_t j = i + 1; j < memoryInsts.size(); ++j) {
+      Instruction* inst1 = memoryInsts[i];
+      Instruction* inst2 = memoryInsts[j];
+      
+      Value* ptr1 = nullptr;
+      Value* ptr2 = nullptr;
+      
+      if (auto* load = dynamic_cast<LoadInst*>(inst1)) {
+        ptr1 = load->getPointer();
+      } else if (auto* store = dynamic_cast<StoreInst*>(inst1)) {
+        ptr1 = store->getPointer();
+      }
+      
+      if (auto* load = dynamic_cast<LoadInst*>(inst2)) {
+        ptr2 = load->getPointer();
+      } else if (auto* store = dynamic_cast<StoreInst*>(inst2)) {
+        ptr2 = store->getPointer();
+      }
+      
+      if (!ptr1 || !ptr2) continue;
+      
+      // 检查是否可能存在别名关系
+      bool mayAlias = false;
+      if (aliasAnalysis) {
+        mayAlias = aliasAnalysis->queryAlias(ptr1, ptr2) != AliasType::NO_ALIAS;
+      } else {
+        mayAlias = (ptr1 != ptr2); // 保守估计
+      }
+      
+      if (mayAlias) {
+        // 创建依赖关系
+        PreciseDependence dep(loop->getLoopDepth());
+        dep.source = inst1;
+        dep.sink = inst2;
+        dep.memoryLocation = ptr1;
+        
+        // 确定依赖类型
+        bool isStore1 = dynamic_cast<StoreInst*>(inst1) != nullptr;
+        bool isStore2 = dynamic_cast<StoreInst*>(inst2) != nullptr;
+        
+        if (isStore1 && !isStore2) {
+          dep.type = DependenceType::TRUE_DEPENDENCE; // Write -> Read (RAW)
+        } else if (!isStore1 && isStore2) {
+          dep.type = DependenceType::ANTI_DEPENDENCE; // Read -> Write (WAR)
+        } else if (isStore1 && isStore2) {
+          dep.type = DependenceType::OUTPUT_DEPENDENCE; // Write -> Write (WAW)
+        } else {
+          continue; // Read -> Read (RAR) - 跳过，不是真正的依赖
+        }
+        
+        // 计算依赖向量
+        dep.dependenceVector = computeAccessDependence(inst1, inst2, loop);
+        
+        // 判断是否允许并行化
+        dep.allowsParallelization = dep.dependenceVector.isLoopIndependent() || 
+                                   (dep.dependenceVector.isKnown && 
+                                    std::all_of(dep.dependenceVector.distances.begin(), 
+                                               dep.dependenceVector.distances.end(),
+                                               [](int d) { return d >= 0; }));
+        
+        dependences.push_back(dep);
+        
+        if (DEBUG && dep.dependenceVector.isKnown) {
+          std::cout << "        Found dependence: " << inst1->getName() 
+                    << " -> " << inst2->getName() << " [";
+          for (size_t k = 0; k < dep.dependenceVector.distances.size(); ++k) {
+            if (k > 0) std::cout << ",";
+            std::cout << dep.dependenceVector.distances[k];
+          }
+          std::cout << "]" << std::endl;
+        }
+      }
+    }
+  }
+  
+  return dependences;
+}
+
+DependenceVector LoopVectorizationPass::computeAccessDependence(Instruction* inst1, Instruction* inst2, Loop* loop) {
+  DependenceVector depVec(loop->getLoopDepth());
+  
+  Value* ptr1 = nullptr;
+  Value* ptr2 = nullptr;
+  
+  if (auto* load = dynamic_cast<LoadInst*>(inst1)) {
+    ptr1 = load->getPointer();
+  } else if (auto* store = dynamic_cast<StoreInst*>(inst1)) {
+    ptr1 = store->getPointer();
+  }
+  
+  if (auto* load = dynamic_cast<LoadInst*>(inst2)) {
+    ptr2 = load->getPointer();
+  } else if (auto* store = dynamic_cast<StoreInst*>(inst2)) {
+    ptr2 = store->getPointer();
+  }
+  
+  if (!ptr1 || !ptr2) return depVec;
+  
+  // 尝试分析仿射关系
+  if (areAccessesAffinelyRelated(ptr1, ptr2, loop)) {
+    auto coeff1 = extractInductionCoefficients(ptr1, loop);
+    auto coeff2 = extractInductionCoefficients(ptr2, loop);
+    
+    if (coeff1.size() == coeff2.size()) {
+      depVec.isKnown = true;
+      depVec.isConstant = true;
+      
+      for (size_t i = 0; i < coeff1.size(); ++i) {
+        depVec.distances[i] = coeff2[i] - coeff1[i];
+      }
+    }
+  }
+  
+  return depVec;
+}
+
+bool LoopVectorizationPass::areAccessesAffinelyRelated(Value* ptr1, Value* ptr2, Loop* loop) {
+  // 简化实现：检查是否都是基于归纳变量的数组访问
+  // 真正的实现需要复杂的仿射关系分析
+  
+  // 检查是否为 GEP 指令
+  auto* gep1 = dynamic_cast<GetElementPtrInst*>(ptr1);
+  auto* gep2 = dynamic_cast<GetElementPtrInst*>(ptr2);
+  
+  if (!gep1 || !gep2) return false;
+  
+  // 检查是否访问同一个数组基址
+  if (gep1->getBasePointer() != gep2->getBasePointer()) return false;
+  
+  // 简化：假设都是仿射的
+  return true;
+}
+
+// ========== 向量化分析实现 (暂时搁置) ==========
+
+VectorizationAnalysis LoopVectorizationPass::analyzeVectorizability(Loop* loop, 
+                                                                   const std::vector<PreciseDependence>& dependences,
+                                                                   LoopCharacteristics* characteristics) {
+  VectorizationAnalysis analysis; // 构造函数已设置为不可向量化
+  
+  if (DEBUG) {
+    std::cout << "    Vectorization analysis: DISABLED (temporarily)" << std::endl;
+  }
+  
+  // 向量化功能暂时搁置，总是返回不可向量化
+  // 这里可以添加一些基本的诊断信息用于日志
+  if (!loop->isInnermost()) {
+    analysis.preventingFactors.push_back("Not innermost loop");
+  }
+  if (loop->getBlocks().size() > 1) {
+    analysis.preventingFactors.push_back("Complex control flow");
+  }
+  if (!dependences.empty()) {
+    analysis.preventingFactors.push_back("Has dependences (not analyzed in detail)");
+  }
+  
+  return analysis;
+}
+
+// ========== 并行化分析实现 ==========
+
+ParallelizationAnalysis LoopVectorizationPass::analyzeParallelizability(Loop* loop,
+                                                                       const std::vector<PreciseDependence>& dependences,
+                                                                       LoopCharacteristics* characteristics) {
+  ParallelizationAnalysis analysis;
+  
+  if (DEBUG) {
+    std::cout << "    Analyzing parallelizability for loop: " << loop->getName() << std::endl;
+    std::cout << "      Found " << dependences.size() << " dependences" << std::endl;
+  }
+  
+  // 按依赖类型分类分析
+  bool hasTrueDependences = false;
+  bool hasAntiDependences = false;
+  bool hasOutputDependences = false;
+  
+  for (const auto& dep : dependences) {
+    switch (dep.type) {
+      case DependenceType::TRUE_DEPENDENCE:
+        hasTrueDependences = true;
+        // 真依赖通常是最难处理的，需要检查是否为归约模式
+        if (dep.isReductionDependence) {
+          analysis.requiresReduction = true;
+          analysis.reductionVariables.insert(dep.memoryLocation);
+        } else {
+          analysis.preventingFactors.push_back("Non-reduction true dependence");
+        }
+        break;
+      case DependenceType::ANTI_DEPENDENCE:
+        hasAntiDependences = true;
+        // 反依赖可以通过变量私有化解决
+        analysis.privatizableVariables.insert(dep.memoryLocation);
+        break;
+      case DependenceType::OUTPUT_DEPENDENCE:
+        hasOutputDependences = true;
+        // 输出依赖可以通过变量私有化或原子操作解决
+        analysis.sharedVariables.insert(dep.memoryLocation);
+        break;
+    }
+  }
+  
+  // 确定并行化类型
+  analysis.parallelType = determineParallelizationType(loop, dependences);
+  
+  // 基于依赖类型评估可并行性
+  if (!hasTrueDependences && !hasOutputDependences) {
+    // 只有反依赖或无依赖，完全可并行
+    analysis.parallelType = ParallelizationAnalysis::EMBARRASSINGLY_PARALLEL;
+    analysis.isParallelizable = true;
+  } else if (analysis.requiresReduction) {
+    // 有归约模式，可以并行但需要特殊处理
+    analysis.parallelType = ParallelizationAnalysis::REDUCTION_PARALLEL;
+    analysis.isParallelizable = true;
+  } else if (hasTrueDependences) {
+    // 有非归约的真依赖，通常不能并行化
+    analysis.isParallelizable = false;
+    analysis.preventingFactors.push_back("Non-reduction loop-carried true dependences");
+  }
+  
+  if (analysis.isParallelizable) {
+    // 进一步分析并行化收益和成本
+    estimateParallelizationBenefit(loop, &analysis, characteristics);
+    analyzeSynchronizationNeeds(loop, &analysis, dependences);
+    analysis.suggestedThreadCount = estimateOptimalThreadCount(loop, characteristics);
+  }
+  
+  if (DEBUG) {
+    std::cout << "      Parallelizable: " << (analysis.isParallelizable ? "YES" : "NO") << std::endl;
+    if (analysis.isParallelizable) {
+      std::cout << "      Type: " << (int)analysis.parallelType << ", Threads: " << analysis.suggestedThreadCount << std::endl;
+    }
+  }
+  
+  return analysis;
+}
+
+bool LoopVectorizationPass::checkParallelizationLegality(Loop* loop, const std::vector<PreciseDependence>& dependences) {
+  // 检查所有依赖是否允许并行化
+  for (const auto& dep : dependences) {
+    if (!dep.allowsParallelization) {
+      return false;
+    }
+  }
+  
+  // 检查是否有无法并行化的操作
+  for (BasicBlock* bb : loop->getBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      // 检查原子操作、同步操作等
+      if (auto* call = dynamic_cast<CallInst*>(inst.get())) {
+        // 简化：假设函数调用需要特殊处理
+        // 在实际实现中，需要分析函数的副作用
+        return false;
+      }
+    }
+  }
+  
+  return true;
+}
+
+int LoopVectorizationPass::estimateOptimalThreadCount(Loop* loop, LoopCharacteristics* characteristics) {
+  // 基于循环特征估计最优线程数
+  if (!characteristics) return 2;
+  
+  // 基于循环体大小和计算密度
+  int baseThreads = 2;
+  
+  if (characteristics->instructionCount > 50) baseThreads = 4;
+  if (characteristics->instructionCount > 200) baseThreads = 8;
+  
+  // 基于计算与内存比率调整
+  if (characteristics->computeToMemoryRatio > 2.0) {
+    baseThreads *= 2; // 计算密集型，可以使用更多线程
+  }
+  
+  return std::min(baseThreads, 16); // 限制最大线程数
+}
+
+// ========== 辅助方法实现 ==========
+
+bool LoopVectorizationPass::isConstantStride(Value* ptr, Loop* loop, int& stride) {
+  // 简化实现：检查是否为常量步长访问
+  stride = 1; // 默认步长
+  
+  auto* gep = dynamic_cast<GetElementPtrInst*>(ptr);
+  if (!gep) return false;
+  
+  // 检查最后一个索引是否为归纳变量 + 常量
+  if (gep->getNumIndices() > 0) {
+    Value* lastIndex = gep->getIndex(gep->getNumIndices() - 1);
+    
+    // 简化：假设是 i 或 i+c 的形式
+    if (auto* binInst = dynamic_cast<BinaryInst*>(lastIndex)) {
+      if (binInst->getKind() == Instruction::kAdd) {
+        // 检查是否为 i + constant
+        if (auto* constInt = dynamic_cast<ConstantInteger*>(binInst->getRhs())) {
+          stride = constInt->getInt();
+          return true;
+        }
+      }
+    }
+    
+    // 默认为步长1的连续访问
+    stride = 1;
+    return true;
+  }
+  
+  return false;
+}
+
+std::vector<int> LoopVectorizationPass::extractInductionCoefficients(Value* ptr, Loop* loop) {
+  // 简化实现：返回默认的仿射系数
+  std::vector<int> coefficients;
+  
+  // 假设是简单的 a[i] 形式，系数为 [0, 1]
+  coefficients.push_back(0); // 常数项
+  coefficients.push_back(1); // 归纳变量系数
+  
+  return coefficients;
+}
+
+// ========== 缺失的方法实现 ==========
+
+ParallelizationAnalysis::ParallelizationType LoopVectorizationPass::determineParallelizationType(
+    Loop* loop, const std::vector<PreciseDependence>& dependences) {
+  
+  // 检查是否有任何依赖
+  if (dependences.empty()) {
+    return ParallelizationAnalysis::EMBARRASSINGLY_PARALLEL;
+  }
+  
+  // 检查是否只有归约模式
+  bool hasReduction = false;
+  bool hasOtherDependences = false;
+  
+  for (const auto& dep : dependences) {
+    if (dep.isReductionDependence) {
+      hasReduction = true;
+    } else if (dep.type == DependenceType::TRUE_DEPENDENCE) {
+      hasOtherDependences = true;
+    }
+  }
+  
+  if (hasReduction && !hasOtherDependences) {
+    return ParallelizationAnalysis::REDUCTION_PARALLEL;
+  } else if (!hasOtherDependences) {
+    return ParallelizationAnalysis::EMBARRASSINGLY_PARALLEL;
+  }
+  
+  return ParallelizationAnalysis::NONE;
+}
+
+void LoopVectorizationPass::analyzeReductionPatterns(Loop* loop, ParallelizationAnalysis* analysis) {
+  // 简化实现：查找常见的归约模式
+  for (BasicBlock* bb : loop->getBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      if (auto* binInst = dynamic_cast<BinaryInst*>(inst.get())) {
+        if (binInst->getKind() == Instruction::kAdd || binInst->getKind() == Instruction::kMul) {
+          // 检查是否为累加/累乘模式
+          Value* lhs = binInst->getLhs();
+          if (hasReductionPattern(lhs, loop)) {
+            analysis->requiresReduction = true;
+            analysis->reductionVariables.insert(lhs);
+          }
+        }
+      }
+    }
+  }
+}
+
+void LoopVectorizationPass::analyzeMemoryAccessPatterns(Loop* loop, ParallelizationAnalysis* analysis, 
+                                                       AliasAnalysisResult* aliasAnalysis) {
+  std::vector<Value*> memoryAccesses;
+  
+  // 收集所有内存访问
+  for (BasicBlock* bb : loop->getBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      if (auto* load = dynamic_cast<LoadInst*>(inst.get())) {
+        memoryAccesses.push_back(load->getPointer());
+      } else if (auto* store = dynamic_cast<StoreInst*>(inst.get())) {
+        memoryAccesses.push_back(store->getPointer());
+      }
+    }
+  }
+  
+  // 分析内存访问独立性
+  bool hasIndependentAccess = true;
+  for (size_t i = 0; i < memoryAccesses.size(); ++i) {
+    for (size_t j = i + 1; j < memoryAccesses.size(); ++j) {
+      if (!isIndependentMemoryAccess(memoryAccesses[i], memoryAccesses[j], loop)) {
+        hasIndependentAccess = false;
+        analysis->hasMemoryConflicts = true;
+      }
+    }
+  }
+  
+  analysis->hasIndependentAccess = hasIndependentAccess;
+}
+
+void LoopVectorizationPass::estimateParallelizationBenefit(Loop* loop, ParallelizationAnalysis* analysis,
+                                                          LoopCharacteristics* characteristics) {
+  if (!analysis->isParallelizable) {
+    analysis->parallelizationBenefit = 0.0;
+    return;
+  }
+  
+  // 基于计算复杂度和并行度计算收益
+  double workComplexity = estimateWorkComplexity(loop);
+  double parallelFraction = 1.0; // 假设完全可并行
+  
+  // 根据依赖调整并行度
+  if (analysis->requiresReduction) {
+    parallelFraction *= 0.8; // 归约降低并行效率
+  }
+  if (analysis->hasMemoryConflicts) {
+    parallelFraction *= 0.6; // 内存冲突降低效率
+  }
+  
+  // Amdahl定律估算
+  double serialFraction = 1.0 - parallelFraction;
+  int threadCount = analysis->suggestedThreadCount;
+  double speedup = 1.0 / (serialFraction + parallelFraction / threadCount);
+  
+  analysis->parallelizationBenefit = std::min((speedup - 1.0) / threadCount, 1.0);
+  
+  // 估算同步和通信开销
+  analysis->synchronizationCost = analysis->requiresBarrier ? 100 : 0;
+  analysis->communicationCost = analysis->sharedVariables.size() * 50;
+}
+
+void LoopVectorizationPass::identifyPrivatizableVariables(Loop* loop, ParallelizationAnalysis* analysis) {
+  // 简化实现：标识循环内定义的变量为可私有化
+  for (BasicBlock* bb : loop->getBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      if (!inst->getType()->isVoid()) {
+        // 如果变量只在循环内使用，可能可以私有化
+        bool onlyUsedInLoop = true;
+        for (auto& use : inst->getUses()) {
+          if (auto* userInst = dynamic_cast<Instruction*>(use->getUser())) {
+            if (!loop->contains(userInst->getParent())) {
+              onlyUsedInLoop = false;
+              break;
+            }
+          }
+        }
+        
+        if (onlyUsedInLoop) {
+          analysis->privatizableVariables.insert(inst.get());
+        }
+      }
+    }
+  }
+}
+
+void LoopVectorizationPass::analyzeSynchronizationNeeds(Loop* loop, ParallelizationAnalysis* analysis,
+                                                       const std::vector<PreciseDependence>& dependences) {
+  // 根据依赖类型确定同步需求
+  for (const auto& dep : dependences) {
+    if (dep.type == DependenceType::OUTPUT_DEPENDENCE) {
+      analysis->requiresBarrier = true;
+      analysis->sharedVariables.insert(dep.memoryLocation);
+    }
+  }
+  
+  // 如果有归约，需要特殊的归约同步
+  if (analysis->requiresReduction) {
+    analysis->requiresBarrier = true;
+  }
+}
+
+bool LoopVectorizationPass::isIndependentMemoryAccess(Value* ptr1, Value* ptr2, Loop* loop) {
+  // 简化实现：基本的独立性检查
+  if (ptr1 == ptr2) return false;
+  
+  // 如果是不同的基址，认为是独立的
+  auto* gep1 = dynamic_cast<GetElementPtrInst*>(ptr1);
+  auto* gep2 = dynamic_cast<GetElementPtrInst*>(ptr2);
+  
+  if (gep1 && gep2) {
+    if (gep1->getBasePointer() != gep2->getBasePointer()) {
+      return true; // 不同的基址
+    }
+    // 相同基址，需要更精细的分析（这里简化为不独立）
+    return false;
+  }
+  
+  return true; // 默认认为独立
+}
+
+double LoopVectorizationPass::estimateWorkComplexity(Loop* loop) {
+  double complexity = 0.0;
+  
+  for (BasicBlock* bb : loop->getBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      // 基于指令类型分配复杂度权重
+      if (auto* binInst = dynamic_cast<BinaryInst*>(inst.get())) {
+        switch (binInst->getKind()) {
+          case Instruction::kAdd:
+          case Instruction::kSub:
+            complexity += 1.0;
+            break;
+          case Instruction::kMul:
+            complexity += 3.0;
+            break;
+          case Instruction::kDiv:
+            complexity += 10.0;
+            break;
+          default:
+            complexity += 2.0;
+        }
+      } else if (dynamic_cast<LoadInst*>(inst.get()) || dynamic_cast<StoreInst*>(inst.get())) {
+        complexity += 2.0; // 内存访问
+      } else {
+        complexity += 1.0; // 其他指令
+      }
+    }
+  }
+  
+  return complexity;
+}
+
+bool LoopVectorizationPass::hasReductionPattern(Value* var, Loop* loop) {
+  // 简化实现：检查是否为简单的累加/累乘模式
+  for (auto& use : var->getUses()) {
+    if (auto* binInst = dynamic_cast<BinaryInst*>(use->getUser())) {
+      if (binInst->getKind() == Instruction::kAdd || binInst->getKind() == Instruction::kMul) {
+        // 检查是否为 var = var op something 的模式
+        if (binInst->getLhs() == var || binInst->getRhs() == var) {
+          return true;
+        }
+      }
+    }
+  }
+  return false;
+}
+
+} // namespace sysy

src/midend/Pass/Analysis/SideEffectAnalysis.cpp

@@ -0,0 +1,413 @@
+#include "SideEffectAnalysis.h"
+#include "AliasAnalysis.h"
+#include "CallGraphAnalysis.h"
+#include "SysYIRPrinter.h"
+#include <iostream>
+
+namespace sysy {
+
+// 副作用分析遍的静态 ID
+void *SysYSideEffectAnalysisPass::ID = (void *)&SysYSideEffectAnalysisPass::ID;
+
+// ======================================================================
+// SideEffectAnalysisResult 类的实现
+// ======================================================================
+
+SideEffectAnalysisResult::SideEffectAnalysisResult() { initializeKnownFunctions(); }
+
+const SideEffectInfo &SideEffectAnalysisResult::getInstructionSideEffect(Instruction *inst) const {
+  auto it = instructionSideEffects.find(inst);
+  if (it != instructionSideEffects.end()) {
+    return it->second;
+  }
+  // 返回默认的无副作用信息
+  static SideEffectInfo noEffect;
+  return noEffect;
+}
+
+const SideEffectInfo &SideEffectAnalysisResult::getFunctionSideEffect(Function *func) const {
+  // 首先检查分析过的用户定义函数
+  auto it = functionSideEffects.find(func);
+  if (it != functionSideEffects.end()) {
+    return it->second;
+  }
+  
+  // 如果没有找到，检查是否为已知的库函数
+  if (func) {
+    std::string funcName = func->getName();
+    const SideEffectInfo *knownInfo = getKnownFunctionSideEffect(funcName);
+    if (knownInfo) {
+      return *knownInfo;
+    }
+  }
+  
+  // 返回默认的无副作用信息
+  static SideEffectInfo noEffect;
+  return noEffect;
+}
+
+void SideEffectAnalysisResult::setInstructionSideEffect(Instruction *inst, const SideEffectInfo &info) {
+  instructionSideEffects[inst] = info;
+}
+
+void SideEffectAnalysisResult::setFunctionSideEffect(Function *func, const SideEffectInfo &info) {
+  functionSideEffects[func] = info;
+}
+
+bool SideEffectAnalysisResult::hasSideEffect(Instruction *inst) const {
+  const auto &info = getInstructionSideEffect(inst);
+  return info.type != SideEffectType::NO_SIDE_EFFECT;
+}
+
+bool SideEffectAnalysisResult::mayModifyMemory(Instruction *inst) const {
+  const auto &info = getInstructionSideEffect(inst);
+  return info.mayModifyMemory;
+}
+
+bool SideEffectAnalysisResult::mayModifyGlobal(Instruction *inst) const {
+  const auto &info = getInstructionSideEffect(inst);
+  return info.mayModifyGlobal;
+}
+
+bool SideEffectAnalysisResult::isPureFunction(Function *func) const {
+  const auto &info = getFunctionSideEffect(func);
+  return info.isPure;
+}
+
+void SideEffectAnalysisResult::initializeKnownFunctions() {
+  // SysY标准库函数的副作用信息
+
+  // I/O函数 - 有副作用
+  SideEffectInfo ioEffect;
+  ioEffect.type = SideEffectType::IO_OPERATION;
+  ioEffect.mayModifyGlobal = true;
+  ioEffect.mayModifyMemory = true;
+  ioEffect.mayCallFunction = true;
+  ioEffect.isPure = false;
+
+  // knownFunctions["printf"] = ioEffect;
+  // knownFunctions["scanf"] = ioEffect;
+  knownFunctions["getint"] = ioEffect;
+  knownFunctions["getch"] = ioEffect;
+  knownFunctions["getfloat"] = ioEffect;
+  knownFunctions["getarray"] = ioEffect;
+  knownFunctions["getfarray"] = ioEffect;
+  knownFunctions["putint"] = ioEffect;
+  knownFunctions["putch"] = ioEffect;
+  knownFunctions["putfloat"] = ioEffect;
+  knownFunctions["putarray"] = ioEffect;
+  knownFunctions["putfarray"] = ioEffect;
+
+  // 时间函数 - 有副作用
+  SideEffectInfo timeEffect;
+  timeEffect.type = SideEffectType::FUNCTION_CALL;
+  timeEffect.mayModifyGlobal = true;
+  timeEffect.mayModifyMemory = false;
+  timeEffect.mayCallFunction = true;
+  timeEffect.isPure = false;
+
+  knownFunctions["_sysy_starttime"] = timeEffect;
+  knownFunctions["_sysy_stoptime"] = timeEffect;
+}
+
+const SideEffectInfo *SideEffectAnalysisResult::getKnownFunctionSideEffect(const std::string &funcName) const {
+  auto it = knownFunctions.find(funcName);
+  return (it != knownFunctions.end()) ? &it->second : nullptr;
+}
+
+// ======================================================================
+// SysYSideEffectAnalysisPass 类的实现
+// ======================================================================
+
+bool SysYSideEffectAnalysisPass::runOnModule(Module *M, AnalysisManager &AM) {
+  if (DEBUG) {
+    std::cout << "Running SideEffect analysis on module" << std::endl;
+  }
+
+  // 创建分析结果（构造函数中已经调用了initializeKnownFunctions）
+  result = std::make_unique<SideEffectAnalysisResult>();
+
+  // 获取调用图分析结果
+  callGraphAnalysis = AM.getAnalysisResult<CallGraphAnalysisResult, CallGraphAnalysisPass>();
+  if (!callGraphAnalysis) {
+    std::cerr << "Warning: CallGraphAnalysis not available, falling back to conservative analysis" << std::endl;
+  }
+
+  // 按拓扑序分析函数，确保被调用函数先于调用者分析
+  if (callGraphAnalysis) {
+    // 使用调用图的拓扑排序结果
+    const auto &topOrder = callGraphAnalysis->getTopologicalOrder();
+
+    // 处理强连通分量（递归函数群）
+    const auto &sccs = callGraphAnalysis->getStronglyConnectedComponents();
+    for (const auto &scc : sccs) {
+      if (scc.size() > 1) {
+        // 多个函数的强连通分量，使用不动点算法
+        analyzeStronglyConnectedComponent(scc, AM);
+      } else {
+        // 单个函数，检查是否自递归
+        Function *func = scc[0];
+        if (callGraphAnalysis->isSelfRecursive(func)) {
+          // 自递归函数也需要不动点算法
+          analyzeStronglyConnectedComponent(scc, AM);
+        } else {
+          // 非递归函数，直接分析
+          SideEffectInfo funcEffect = analyzeFunction(func, AM);
+          result->setFunctionSideEffect(func, funcEffect);
+        }
+      }
+    }
+  } else {
+    // 没有调用图，保守地分析每个函数
+    for (auto &pair : M->getFunctions()) {
+      Function *func = pair.second.get();
+      SideEffectInfo funcEffect = analyzeFunction(func, AM);
+      result->setFunctionSideEffect(func, funcEffect);
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "---- Side Effect Analysis Results for Module ----\n";
+    for (auto &pair : M->getFunctions()) {
+      Function *func = pair.second.get();
+      const auto &funcInfo = result->getFunctionSideEffect(func);
+
+      std::cout << "Function " << func->getName() << ": ";
+      switch (funcInfo.type) {
+      case SideEffectType::NO_SIDE_EFFECT:
+        std::cout << "No Side Effect";
+        break;
+      case SideEffectType::MEMORY_WRITE:
+        std::cout << "Memory Write";
+        break;
+      case SideEffectType::FUNCTION_CALL:
+        std::cout << "Function Call";
+        break;
+      case SideEffectType::IO_OPERATION:
+        std::cout << "I/O Operation";
+        break;
+      case SideEffectType::UNKNOWN:
+        std::cout << "Unknown";
+        break;
+      }
+      std::cout << " (Pure: " << (funcInfo.isPure ? "Yes" : "No")
+                << ", Modifies Global: " << (funcInfo.mayModifyGlobal ? "Yes" : "No") << ")\n";
+    }
+    std::cout << "--------------------------------------------------\n";
+  }
+
+  return false; // Analysis passes return false since they don't modify the IR
+}
+
+std::unique_ptr<AnalysisResultBase> SysYSideEffectAnalysisPass::getResult() { return std::move(result); }
+
+SideEffectInfo SysYSideEffectAnalysisPass::analyzeFunction(Function *func, AnalysisManager &AM) {
+  SideEffectInfo functionSideEffect;
+
+  // 为每个指令分析副作用
+  for (auto &BB : func->getBasicBlocks()) {
+    for (auto &I : BB->getInstructions_Range()) {
+      Instruction *inst = I.get();
+      SideEffectInfo instEffect = analyzeInstruction(inst, func, AM);
+
+      // 记录指令的副作用信息
+      result->setInstructionSideEffect(inst, instEffect);
+
+      // 合并到函数级别的副作用信息中
+      functionSideEffect = functionSideEffect.merge(instEffect);
+    }
+  }
+
+  return functionSideEffect;
+}
+
+void SysYSideEffectAnalysisPass::analyzeStronglyConnectedComponent(const std::vector<Function *> &scc,
+                                                                   AnalysisManager &AM) {
+  // 使用不动点算法处理递归函数群
+  std::unordered_map<Function *, SideEffectInfo> currentEffects;
+  std::unordered_map<Function *, SideEffectInfo> previousEffects;
+
+  // 初始化：所有函数都假设为纯函数
+  for (Function *func : scc) {
+    SideEffectInfo initialEffect;
+    initialEffect.isPure = true;
+    currentEffects[func] = initialEffect;
+    result->setFunctionSideEffect(func, initialEffect);
+  }
+
+  bool converged = false;
+  int iterations = 0;
+  const int maxIterations = 10; // 防止无限循环
+
+  while (!converged && iterations < maxIterations) {
+    previousEffects = currentEffects;
+
+    // 重新分析每个函数
+    for (Function *func : scc) {
+      SideEffectInfo newEffect = analyzeFunction(func, AM);
+      currentEffects[func] = newEffect;
+      result->setFunctionSideEffect(func, newEffect);
+    }
+
+    // 检查是否收敛
+    converged = hasConverged(previousEffects, currentEffects);
+    iterations++;
+  }
+
+  if (iterations >= maxIterations) {
+    std::cerr << "Warning: SideEffect analysis did not converge for SCC after " << maxIterations << " iterations"
+              << std::endl;
+  }
+}
+
+bool SysYSideEffectAnalysisPass::hasConverged(const std::unordered_map<Function *, SideEffectInfo> &oldEffects,
+                                              const std::unordered_map<Function *, SideEffectInfo> &newEffects) const {
+  for (const auto &pair : oldEffects) {
+    Function *func = pair.first;
+    const SideEffectInfo &oldEffect = pair.second;
+
+    auto it = newEffects.find(func);
+    if (it == newEffects.end()) {
+      return false; // 函数不存在于新结果中
+    }
+
+    const SideEffectInfo &newEffect = it->second;
+
+    // 比较关键属性是否相同
+    if (oldEffect.type != newEffect.type || oldEffect.mayModifyGlobal != newEffect.mayModifyGlobal ||
+        oldEffect.mayModifyMemory != newEffect.mayModifyMemory ||
+        oldEffect.mayCallFunction != newEffect.mayCallFunction || oldEffect.isPure != newEffect.isPure) {
+      return false;
+    }
+  }
+
+  return true;
+}
+
+SideEffectInfo SysYSideEffectAnalysisPass::analyzeInstruction(Instruction *inst, Function *currentFunc,
+                                                              AnalysisManager &AM) {
+  SideEffectInfo info;
+
+  // 根据指令类型进行分析
+  if (inst->isCall()) {
+    return analyzeCallInstruction(static_cast<CallInst *>(inst), currentFunc, AM);
+  } else if (inst->isStore()) {
+    return analyzeStoreInstruction(static_cast<StoreInst *>(inst), currentFunc, AM);
+  } else if (inst->isMemset()) {
+    return analyzeMemsetInstruction(static_cast<MemsetInst *>(inst), currentFunc, AM);
+  } else if (inst->isBranch() || inst->isReturn()) {
+    // 控制流指令无副作用，但必须保留
+    info.type = SideEffectType::NO_SIDE_EFFECT;
+    info.isPure = true;
+  } else {
+    // 其他指令（算术、逻辑、比较等）通常无副作用
+    info.type = SideEffectType::NO_SIDE_EFFECT;
+    info.isPure = true;
+  }
+
+  return info;
+}
+
+SideEffectInfo SysYSideEffectAnalysisPass::analyzeCallInstruction(CallInst *call, Function *currentFunc,
+                                                                  AnalysisManager &AM) {
+  SideEffectInfo info;
+
+  // 获取被调用的函数
+  Function *calledFunc = call->getCallee();
+  if (!calledFunc) {
+    // 间接调用，保守处理
+    info.type = SideEffectType::UNKNOWN;
+    info.mayModifyGlobal = true;
+    info.mayModifyMemory = true;
+    info.mayCallFunction = true;
+    info.isPure = false;
+    return info;
+  }
+
+  std::string funcName = calledFunc->getName();
+
+  // 检查是否为已知的标准库函数
+  const SideEffectInfo *knownInfo = result->getKnownFunctionSideEffect(funcName);
+  if (knownInfo) {
+    return *knownInfo;
+  }
+
+  // 利用调用图分析结果进行精确分析
+  if (callGraphAnalysis) {
+    // 检查被调用函数是否已分析过
+    const SideEffectInfo &funcEffect = result->getFunctionSideEffect(calledFunc);
+    if (funcEffect.type != SideEffectType::NO_SIDE_EFFECT || !funcEffect.isPure) {
+      return funcEffect;
+    }
+
+    // 检查递归调用
+    if (callGraphAnalysis->isRecursive(calledFunc)) {
+      // 递归函数保守处理（在不动点算法中会精确分析）
+      info.type = SideEffectType::FUNCTION_CALL;
+      info.mayModifyGlobal = true;
+      info.mayModifyMemory = true;
+      info.mayCallFunction = true;
+      info.isPure = false;
+      return info;
+    }
+  }
+
+  // 对于未分析的用户函数，保守处理
+  info.type = SideEffectType::FUNCTION_CALL;
+  info.mayModifyGlobal = true;
+  info.mayModifyMemory = true;
+  info.mayCallFunction = true;
+  info.isPure = false;
+
+  return info;
+}
+
+SideEffectInfo SysYSideEffectAnalysisPass::analyzeStoreInstruction(StoreInst *store, Function *currentFunc,
+                                                                   AnalysisManager &AM) {
+  SideEffectInfo info;
+  info.type = SideEffectType::MEMORY_WRITE;
+  info.mayModifyMemory = true;
+  info.isPure = false;
+
+  // 获取函数的别名分析结果
+  AliasAnalysisResult *aliasAnalysis = AM.getAnalysisResult<AliasAnalysisResult, SysYAliasAnalysisPass>(currentFunc);
+  if (aliasAnalysis) {
+    Value *storePtr = store->getPointer();
+
+    // 如果存储到全局变量或可能别名的位置，则可能修改全局状态
+    if (!aliasAnalysis->isLocalArray(storePtr)) {
+      info.mayModifyGlobal = true;
+    }
+  } else {
+    // 没有别名分析结果，保守处理
+    info.mayModifyGlobal = true;
+  }
+
+  return info;
+}
+
+SideEffectInfo SysYSideEffectAnalysisPass::analyzeMemsetInstruction(MemsetInst *memset, Function *currentFunc,
+                                                                    AnalysisManager &AM) {
+  SideEffectInfo info;
+  info.type = SideEffectType::MEMORY_WRITE;
+  info.mayModifyMemory = true;
+  info.isPure = false;
+
+  // 获取函数的别名分析结果
+  AliasAnalysisResult *aliasAnalysis = AM.getAnalysisResult<AliasAnalysisResult, SysYAliasAnalysisPass>(currentFunc);
+  if (aliasAnalysis) {
+    Value *memsetPtr = memset->getPointer();
+
+    // 如果memset操作全局变量或可能别名的位置，则可能修改全局状态
+    if (!aliasAnalysis->isLocalArray(memsetPtr)) {
+      info.mayModifyGlobal = true;
+    }
+  } else {
+    // 没有别名分析结果，保守处理
+    info.mayModifyGlobal = true;
+  }
+
+  return info;
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/BuildCFG.cpp

@@ -0,0 +1,79 @@
+#include "BuildCFG.h"
+#include "Dom.h"
+#include "Liveness.h"
+#include <iostream>
+#include <queue>
+#include <set>
+
+namespace sysy {
+
+void *BuildCFG::ID = (void *)&BuildCFG::ID; // 定义唯一的 Pass ID
+
+// 声明Pass的分析使用
+void BuildCFG::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const {
+  // BuildCFG不依赖其他分析
+  // analysisDependencies.insert(&DominatorTreeAnalysisPass::ID); // 错误的例子
+
+  // BuildCFG会使所有依赖于CFG的分析结果失效，所以它必须声明这些失效
+  analysisInvalidations.insert(&DominatorTreeAnalysisPass::ID);
+  analysisInvalidations.insert(&LivenessAnalysisPass::ID);
+}
+
+bool BuildCFG::runOnFunction(Function *F, AnalysisManager &AM) {
+  if (DEBUG) {
+    std::cout << "Running BuildCFG pass on function: " << F->getName() << std::endl;
+  }
+
+  bool changed = false;
+
+  // 1. 清空所有基本块的前驱和后继列表
+  for (auto &bb : F->getBasicBlocks()) {
+    bb->clearPredecessors();
+    bb->clearSuccessors();
+  }
+
+  // 2. 遍历每个基本块，重建CFG
+  for (auto &bb : F->getBasicBlocks()) {
+    // 获取基本块的最后一条指令
+    auto &inst = *bb->terminator();
+    Instruction *termInst = inst.get();
+    // 确保基本块有终结指令
+    if (!termInst) {
+      continue;
+    }
+
+    // 根据终结指令类型，建立前驱后继关系
+    if (termInst->isBranch()) {
+      // 无条件跳转
+      if (termInst->isUnconditional()) {
+        auto brInst = dynamic_cast<UncondBrInst *>(termInst);
+        BasicBlock *succ = dynamic_cast<BasicBlock *>(brInst->getBlock());
+        assert(succ && "Branch instruction's target must be a BasicBlock");
+        bb->addSuccessor(succ);
+        succ->addPredecessor(bb.get());
+        changed = true;
+
+        // 条件跳转
+      } else if (termInst->isConditional()) {
+        auto brInst = dynamic_cast<CondBrInst *>(termInst);
+        BasicBlock *trueSucc = dynamic_cast<BasicBlock *>(brInst->getThenBlock());
+        BasicBlock *falseSucc = dynamic_cast<BasicBlock *>(brInst->getElseBlock());
+
+        assert(trueSucc && falseSucc && "Branch instruction's targets must be BasicBlocks");
+
+        bb->addSuccessor(trueSucc);
+        trueSucc->addPredecessor(bb.get());
+        bb->addSuccessor(falseSucc);
+        falseSucc->addPredecessor(bb.get());
+        changed = true;
+      }
+    } else if (auto retInst = dynamic_cast<ReturnInst *>(termInst)) {
+      // RetInst没有后继，无需处理
+      // ...
+    }
+  }
+
+  return changed;
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/DCE.cpp

@@ -1,9 +1,9 @@
-#include "DCE.h"            // 包含DCE遍的头文件
-#include "IR.h"             // 包含IR相关的定义
-#include "SysYIROptUtils.h" // 包含SysY IR优化工具类的定义
-#include <cassert>          // 用于断言
-#include <iostream>         // 用于调试输出
-#include <set>              // 包含set，虽然DCEContext内部用unordered_set，但这里保留
+#include "DCE.h"
+#include "SysYIROptUtils.h"
+#include "SideEffectAnalysis.h"
+#include <cassert>          
+#include <iostream>         
+#include <set>              
 
 namespace sysy {
 
@@ -17,10 +17,26 @@ void *DCE::ID = (void *)&DCE::ID;
 
 // DCEContext 的 run 方法实现
 void DCEContext::run(Function *func, AnalysisManager *AM, bool &changed) {
+  // 获取别名分析结果
+  if (AM) {
+    aliasAnalysis = AM->getAnalysisResult<AliasAnalysisResult, SysYAliasAnalysisPass>(func);
+    // 获取副作用分析结果（Module级别）
+    sideEffectAnalysis = AM->getAnalysisResult<SideEffectAnalysisResult, SysYSideEffectAnalysisPass>();
+    
+    if (DEBUG) {
+      if (aliasAnalysis) {
+        std::cout << "DCE: Using alias analysis results" << std::endl;
+      }
+      if (sideEffectAnalysis) {
+        std::cout << "DCE: Using side effect analysis results" << std::endl;
+      }
+    }
+  }
+  
   // 清空活跃指令集合，确保每次运行都是新的状态
   alive_insts.clear();
 
-  // 第一次遍历：扫描所有指令，识别“天然活跃”的指令并将其及其依赖标记为活跃
+  // 第一次遍历：扫描所有指令，识别"天然活跃"的指令并将其及其依赖标记为活跃
   // 使用 func->getBasicBlocks() 获取基本块列表，保留用户风格
   auto basicBlocks = func->getBasicBlocks();
   for (auto &basicBlock : basicBlocks) {
@@ -51,7 +67,7 @@ void DCEContext::run(Function *func, AnalysisManager *AM, bool &changed) {
       // 如果指令不在活跃集合中，则删除它。
       // 分支和返回指令由 isAlive 处理，并会被保留。
       if (alive_insts.count(currentInst) == 0) {
-        instIter =  SysYIROptUtils::usedelete(instIter); // 删除后返回下一个迭代器
+        instIter = SysYIROptUtils::usedelete(instIter); // 删除后返回下一个迭代器
         changed = true; // 标记 IR 已被修改
       } else {
         ++instIter; // 指令活跃，移动到下一个
@@ -60,20 +76,58 @@ void DCEContext::run(Function *func, AnalysisManager *AM, bool &changed) {
   }
 }
 
-// 判断指令是否是“天然活跃”的实现
+// 判断指令是否是"天然活跃"的实现
 // 只有具有副作用的指令（如存储、函数调用、原子操作）
 // 和控制流指令（如分支、返回）是天然活跃的。
 bool DCEContext::isAlive(Instruction *inst) {
-  // TODO: 后续程序并发考虑原子操作
-  // 其结果不被其他指令使用的指令（例如 StoreInst, BranchInst, ReturnInst）。
-  // dynamic_cast<ir::CallInst>(inst) 检查是否是函数调用指令，
-  // 函数调用通常有副作用。
-  // 终止指令 (BranchInst, ReturnInst) 必须是活跃的，因为它控制了程序的执行流程。
-  // 保留用户提供的 isAlive 逻辑
-  bool isBranchOrReturn = inst->isBranch() || inst->isReturn();
-  bool isCall = inst->isCall();
-  bool isStoreOrMemset = inst->isStore() || inst->isMemset(); 
-  return isBranchOrReturn || isCall || isStoreOrMemset;
+  // 终止指令 (BranchInst, ReturnInst) 必须是活跃的，因为它控制了程序的执行流程
+  if (inst->isBranch() || inst->isReturn()) {
+    return true;
+  }
+  
+  // 使用副作用分析来判断指令是否有副作用
+  if (sideEffectAnalysis && sideEffectAnalysis->hasSideEffect(inst)) {
+    return true;
+  }
+  
+  // 特殊处理Store指令：使用别名分析进行更精确的判断
+  if (inst->isStore()) {
+    auto* storeInst = static_cast<StoreInst*>(inst);
+    return mayHaveSideEffect(storeInst);
+  }
+  
+  // 特殊处理Memset指令：总是保留（因为它修改内存）
+  if (inst->isMemset()) {
+    return true;
+  }
+  
+  // 函数调用指令：总是保留（可能有未知副作用）
+  if (inst->isCall()) {
+    return true;
+  }
+  
+  // 其他指令（算术、逻辑、Load等）：无副作用，可以删除
+  return false;
+}
+
+// 检查Store指令是否可能有副作用（通过别名分析）
+bool DCEContext::mayHaveSideEffect(StoreInst* store) {
+  if (!aliasAnalysis) {
+    // 没有别名分析结果时，保守地认为所有store都有副作用
+    return true;
+  }
+  
+  Value* storePtr = store->getPointer();
+  
+  // 如果是对本地数组的存储且访问模式是常量，可能可以安全删除
+  if (aliasAnalysis->isLocalArray(storePtr)) {
+    // 检查是否有其他指令可能读取这个位置
+    // 这里需要更复杂的活性分析，暂时保守处理
+    return true; // 保守地保留所有本地数组的存储
+  }
+  
+  // 对全局变量、函数参数等的存储总是有副作用
+  return true;
 }
 
 // 递归地将活跃指令及其依赖加入到 alive_insts 集合中
@@ -102,7 +156,6 @@ void DCEContext::addAlive(Instruction *inst) {
 
 // DCE 遍的 runOnFunction 方法实现
 bool DCE::runOnFunction(Function *func, AnalysisManager &AM) {
-
   DCEContext ctx;
   bool changed = false;
   ctx.run(func, &AM, changed); // 运行 DCE 优化
@@ -120,7 +173,11 @@ bool DCE::runOnFunction(Function *func, AnalysisManager &AM) {
 
 // 声明DCE遍的分析依赖和失效信息
 void DCE::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const {
-  // DCE不依赖特定的分析结果，它通过遍历和副作用判断来工作。
+  // DCE依赖别名分析来更精确地判断Store指令的副作用
+  analysisDependencies.insert(&SysYAliasAnalysisPass::ID);
+  
+  // DCE依赖副作用分析来判断指令是否有副作用
+  analysisDependencies.insert(&SysYSideEffectAnalysisPass::ID);
 
   // DCE会删除指令，这会影响许多分析结果。
   // 至少，它会影响活跃性分析、支配树、控制流图（如果删除导致基本块为空并被合并）。

src/midend/Pass/Optimize/GVN.cpp

@@ -0,0 +1,492 @@
+#include "GVN.h"
+#include "Dom.h"
+#include "SysYIROptUtils.h"
+#include <algorithm>
+#include <cassert>
+#include <iostream>
+#include <unordered_map>
+#include <unordered_set>
+
+extern int DEBUG;
+
+namespace sysy {
+
+// GVN 遍的静态 ID
+void *GVN::ID = (void *)&GVN::ID;
+
+// ======================================================================
+// GVN 类的实现
+// ======================================================================
+
+bool GVN::runOnFunction(Function *func, AnalysisManager &AM) {
+  if (func->getBasicBlocks().empty()) {
+    return false;
+  }
+
+  if (DEBUG) {
+    std::cout << "\n=== Running GVN on function: " << func->getName() << " ===" << std::endl;
+  }
+
+  bool changed = false;
+  GVNContext context;
+  context.run(func, &AM, changed);
+
+  if (DEBUG) {
+    if (changed) {
+      std::cout << "GVN: Function " << func->getName() << " was modified" << std::endl;
+    } else {
+      std::cout << "GVN: Function " << func->getName() << " was not modified" << std::endl;
+    }
+    std::cout << "=== GVN completed for function: " << func->getName() << " ===" << std::endl;
+  }
+
+  return changed;
+}
+
+void GVN::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const {
+  // GVN依赖以下分析：
+  // 1. 支配树分析 - 用于检查指令的支配关系，确保替换的安全性
+  analysisDependencies.insert(&DominatorTreeAnalysisPass::ID);
+  
+  // 2. 副作用分析 - 用于判断函数调用是否可以进行GVN
+  analysisDependencies.insert(&SysYSideEffectAnalysisPass::ID);
+
+  // GVN不会使任何分析失效，因为：
+  // - GVN只删除冗余计算，不改变CFG结构
+  // - GVN不修改程序的语义，只是消除重复计算
+  // - 支配关系保持不变
+  // - 副作用分析结果保持不变
+  // analysisInvalidations 保持为空
+  
+  if (DEBUG) {
+    std::cout << "GVN: Declared analysis dependencies (DominatorTree, SideEffectAnalysis)" << std::endl;
+  }
+}
+
+// ======================================================================
+// GVNContext 类的实现 - 重构版本
+// ======================================================================
+
+// 简单的表达式哈希结构
+struct ExpressionKey {
+  enum Type { BINARY, UNARY, LOAD, GEP, CALL } type;
+  int opcode;
+  std::vector<Value*> operands;
+  Type* resultType;
+  
+  bool operator==(const ExpressionKey& other) const {
+    return type == other.type && opcode == other.opcode && 
+           operands == other.operands && resultType == other.resultType;
+  }
+};
+
+struct ExpressionKeyHash {
+  size_t operator()(const ExpressionKey& key) const {
+    size_t hash = std::hash<int>()(static_cast<int>(key.type)) ^ 
+                  std::hash<int>()(key.opcode);
+    for (auto op : key.operands) {
+      hash ^= std::hash<Value*>()(op) + 0x9e3779b9 + (hash << 6) + (hash >> 2);
+    }
+    return hash;
+  }
+};
+
+void GVNContext::run(Function *func, AnalysisManager *AM, bool &changed) {
+  if (DEBUG) {
+    std::cout << "  Starting GVN analysis for function: " << func->getName() << std::endl;
+  }
+
+  // 获取分析结果
+  if (AM) {
+    domTree = AM->getAnalysisResult<DominatorTree, DominatorTreeAnalysisPass>(func);
+    sideEffectAnalysis = AM->getAnalysisResult<SideEffectAnalysisResult, SysYSideEffectAnalysisPass>();
+
+    if (DEBUG) {
+      if (domTree) {
+        std::cout << "    GVN: Using dominator tree analysis" << std::endl;
+      } else {
+        std::cout << "    GVN: Warning - dominator tree analysis not available" << std::endl;
+      }
+      if (sideEffectAnalysis) {
+        std::cout << "    GVN: Using side effect analysis" << std::endl;
+      } else {
+        std::cout << "    GVN: Warning - side effect analysis not available" << std::endl;
+      }
+    }
+  }
+
+  // 清空状态
+  valueToNumber.clear();
+  numberToValue.clear();
+  expressionToNumber.clear();
+  nextValueNumber = 1;
+  visited.clear();
+  rpoBlocks.clear();
+  needRemove.clear();
+
+  // 计算逆后序遍历
+  computeRPO(func);
+
+  if (DEBUG) {
+    std::cout << "    Computed RPO with " << rpoBlocks.size() << " blocks" << std::endl;
+  }
+
+  // 按逆后序遍历基本块进行GVN
+  int blockCount = 0;
+  for (auto bb : rpoBlocks) {
+    if (DEBUG) {
+      std::cout << "    Processing block " << ++blockCount << "/" << rpoBlocks.size() 
+                << ": " << bb->getName() << std::endl;
+    }
+    
+    processBasicBlock(bb, changed);
+  }
+
+  if (DEBUG) {
+    std::cout << "    Found " << needRemove.size() << " redundant instructions to remove" << std::endl;
+  }
+
+  // 删除冗余指令
+  eliminateRedundantInstructions(changed);
+
+  if (DEBUG) {
+    std::cout << "  GVN analysis completed for function: " << func->getName() << std::endl;
+    std::cout << "    Total values numbered: " << valueToNumber.size() << std::endl;
+    std::cout << "    Instructions eliminated: " << needRemove.size() << std::endl;
+  }
+}
+
+void GVNContext::computeRPO(Function *func) {
+  rpoBlocks.clear();
+  visited.clear();
+
+  auto entry = func->getEntryBlock();
+  if (entry) {
+    dfs(entry);
+    std::reverse(rpoBlocks.begin(), rpoBlocks.end());
+  }
+}
+
+void GVNContext::dfs(BasicBlock *bb) {
+  if (!bb || visited.count(bb)) {
+    return;
+  }
+
+  visited.insert(bb);
+
+  // 访问所有后继基本块
+  for (auto succ : bb->getSuccessors()) {
+    if (visited.find(succ) == visited.end()) {
+      dfs(succ);
+    }
+  }
+
+  rpoBlocks.push_back(bb);
+}
+
+unsigned GVNContext::getValueNumber(Value* value) {
+  // 如果已经有值编号，直接返回
+  auto it = valueToNumber.find(value);
+  if (it != valueToNumber.end()) {
+    return it->second;
+  }
+  
+  // 为新值分配编号
+  return assignValueNumber(value);
+}
+
+unsigned GVNContext::assignValueNumber(Value* value) {
+  unsigned number = nextValueNumber++;
+  valueToNumber[value] = number;
+  numberToValue[number] = value;
+  
+  if (DEBUG >= 2) {
+    std::cout << "            Assigned value number " << number 
+              << " to " << value->getName() << std::endl;
+  }
+  
+  return number;
+}
+
+void GVNContext::processBasicBlock(BasicBlock* bb, bool& changed) {
+  int instCount = 0;
+  for (auto &instPtr : bb->getInstructions()) {
+    if (DEBUG) {
+      std::cout << "      Processing instruction " << ++instCount 
+                << ": " << instPtr->getName() << std::endl;
+    }
+    
+    if (processInstruction(instPtr.get())) {
+      changed = true;
+    }
+  }
+}
+
+bool GVNContext::processInstruction(Instruction* inst) {
+  // 跳过分支指令和其他不可优化的指令
+  if (inst->isBranch() || dynamic_cast<ReturnInst*>(inst) || 
+      dynamic_cast<AllocaInst*>(inst) || dynamic_cast<StoreInst*>(inst)) {
+    
+    // 如果是store指令，需要使相关的内存值失效
+    if (auto store = dynamic_cast<StoreInst*>(inst)) {
+      invalidateMemoryValues(store);
+    }
+    
+    // 为这些指令分配值编号但不尝试优化
+    getValueNumber(inst);
+    return false;
+  }
+  
+  if (DEBUG) {
+    std::cout << "        Processing optimizable instruction: " << inst->getName() 
+              << " (kind: " << static_cast<int>(inst->getKind()) << ")" << std::endl;
+  }
+  
+  // 构建表达式键
+  std::string exprKey = buildExpressionKey(inst);
+  if (exprKey.empty()) {
+    // 不可优化的指令，只分配值编号
+    getValueNumber(inst);
+    return false;
+  }
+  
+  if (DEBUG >= 2) {
+    std::cout << "          Expression key: " << exprKey << std::endl;
+  }
+  
+  // 查找已存在的等价值
+  Value* existing = findExistingValue(exprKey, inst);
+  if (existing && existing != inst) {
+    // 检查支配关系
+    if (auto existingInst = dynamic_cast<Instruction*>(existing)) {
+      if (dominates(existingInst, inst)) {
+        if (DEBUG) {
+          std::cout << "        GVN: Replacing " << inst->getName() 
+                    << " with existing " << existing->getName() << std::endl;
+        }
+        
+        // 用已存在的值替换当前指令
+        inst->replaceAllUsesWith(existing);
+        needRemove.insert(inst);
+        
+        // 将当前指令的值编号指向已存在的值
+        unsigned existingNumber = getValueNumber(existing);
+        valueToNumber[inst] = existingNumber;
+        
+        return true;
+      } else {
+        if (DEBUG) {
+          std::cout << "          Found equivalent but dominance check failed" << std::endl;
+        }
+      }
+    }
+  }
+  
+  // 没有找到等价值，为这个表达式分配新的值编号
+  unsigned number = assignValueNumber(inst);
+  expressionToNumber[exprKey] = number;
+  
+  if (DEBUG) {
+    std::cout << "        Instruction " << inst->getName() << " is unique" << std::endl;
+  }
+  
+  return false;
+}
+
+std::string GVNContext::buildExpressionKey(Instruction* inst) {
+  std::ostringstream oss;
+  
+  if (auto binary = dynamic_cast<BinaryInst*>(inst)) {
+    oss << "binary_" << static_cast<int>(binary->getKind()) << "_";
+    oss << getValueNumber(binary->getLhs()) << "_" << getValueNumber(binary->getRhs());
+    
+    // 对于可交换操作，确保操作数顺序一致
+    if (binary->isCommutative()) {
+      unsigned lhsNum = getValueNumber(binary->getLhs());
+      unsigned rhsNum = getValueNumber(binary->getRhs());
+      if (lhsNum > rhsNum) {
+        oss.str("");
+        oss << "binary_" << static_cast<int>(binary->getKind()) << "_";
+        oss << rhsNum << "_" << lhsNum;
+      }
+    }
+  } else if (auto unary = dynamic_cast<UnaryInst*>(inst)) {
+    oss << "unary_" << static_cast<int>(unary->getKind()) << "_";
+    oss << getValueNumber(unary->getOperand());
+  } else if (auto gep = dynamic_cast<GetElementPtrInst*>(inst)) {
+    oss << "gep_" << getValueNumber(gep->getBasePointer());
+    for (unsigned i = 0; i < gep->getNumIndices(); ++i) {
+      oss << "_" << getValueNumber(gep->getIndex(i));
+    }
+  } else if (auto load = dynamic_cast<LoadInst*>(inst)) {
+    oss << "load_" << getValueNumber(load->getPointer());
+    oss << "_" << reinterpret_cast<uintptr_t>(load->getType()); // 类型区分
+  } else if (auto call = dynamic_cast<CallInst*>(inst)) {
+    // 只为无副作用的函数调用建立表达式
+    if (sideEffectAnalysis && sideEffectAnalysis->isPureFunction(call->getCallee())) {
+      oss << "call_" << call->getCallee()->getName();
+      for (size_t i = 1; i < call->getNumOperands(); ++i) { // 跳过函数指针
+        oss << "_" << getValueNumber(call->getOperand(i));
+      }
+    } else {
+      return ""; // 有副作用的函数调用不可优化
+    }
+  } else {
+    return ""; // 不支持的指令类型
+  }
+  
+  return oss.str();
+}
+
+Value* GVNContext::findExistingValue(const std::string& exprKey, Instruction* inst) {
+  auto it = expressionToNumber.find(exprKey);
+  if (it != expressionToNumber.end()) {
+    unsigned number = it->second;
+    auto valueIt = numberToValue.find(number);
+    if (valueIt != numberToValue.end()) {
+      Value* existing = valueIt->second;
+      
+      // 对于load指令，需要额外检查内存安全性
+      if (auto loadInst = dynamic_cast<LoadInst*>(inst)) {
+        if (auto existingLoad = dynamic_cast<LoadInst*>(existing)) {
+          if (!isMemorySafe(existingLoad, loadInst)) {
+            return nullptr;
+          }
+        }
+      }
+      
+      return existing;
+    }
+  }
+  return nullptr;
+}
+
+bool GVNContext::dominates(Instruction* a, Instruction* b) {
+  auto aBB = a->getParent();
+  auto bBB = b->getParent();
+  
+  // 同一基本块内的情况
+  if (aBB == bBB) {
+    auto &insts = aBB->getInstructions();
+    auto aIt = std::find_if(insts.begin(), insts.end(), 
+                           [a](const auto &ptr) { return ptr.get() == a; });
+    auto bIt = std::find_if(insts.begin(), insts.end(),
+                           [b](const auto &ptr) { return ptr.get() == b; });
+    
+    if (aIt == insts.end() || bIt == insts.end()) {
+      return false;
+    }
+    
+    return std::distance(insts.begin(), aIt) < std::distance(insts.begin(), bIt);
+  }
+  
+  // 不同基本块的情况，使用支配树
+  if (domTree) {
+    auto dominators = domTree->getDominators(bBB);
+    return dominators && dominators->count(aBB);
+  }
+  
+  return false; // 保守做法
+}
+
+bool GVNContext::isMemorySafe(LoadInst* earlierLoad, LoadInst* laterLoad) {
+  // 检查两个load是否访问相同的内存位置
+  unsigned earlierPtr = getValueNumber(earlierLoad->getPointer());
+  unsigned laterPtr = getValueNumber(laterLoad->getPointer());
+  
+  if (earlierPtr != laterPtr) {
+    return false; // 不同的内存位置
+  }
+  
+  // 检查类型是否匹配
+  if (earlierLoad->getType() != laterLoad->getType()) {
+    return false;
+  }
+  
+  // 简单情况：如果在同一个基本块且没有中间的store，则安全
+  auto earlierBB = earlierLoad->getParent();
+  auto laterBB = laterLoad->getParent();
+  
+  if (earlierBB != laterBB) {
+    // 跨基本块的情况需要更复杂的分析，暂时保守处理
+    return false;
+  }
+  
+  // 同一基本块内检查是否有中间的store
+  auto &insts = earlierBB->getInstructions();
+  auto earlierIt = std::find_if(insts.begin(), insts.end(),
+                               [earlierLoad](const auto &ptr) { return ptr.get() == earlierLoad; });
+  auto laterIt = std::find_if(insts.begin(), insts.end(),
+                              [laterLoad](const auto &ptr) { return ptr.get() == laterLoad; });
+  
+  if (earlierIt == insts.end() || laterIt == insts.end()) {
+    return false;
+  }
+  
+  // 确保earlierLoad真的在laterLoad之前
+  if (std::distance(insts.begin(), earlierIt) >= std::distance(insts.begin(), laterIt)) {
+    return false;
+  }
+  
+  // 检查中间是否有store指令修改了相同的内存位置
+  for (auto it = std::next(earlierIt); it != laterIt; ++it) {
+    if (auto store = dynamic_cast<StoreInst*>(it->get())) {
+      unsigned storePtr = getValueNumber(store->getPointer());
+      if (storePtr == earlierPtr) {
+        return false; // 找到中间的store
+      }
+    }
+    
+    // 检查函数调用是否可能修改内存
+    if (auto call = dynamic_cast<CallInst*>(it->get())) {
+      if (sideEffectAnalysis && !sideEffectAnalysis->isPureFunction(call->getCallee())) {
+        // 保守处理：有副作用的函数可能修改内存
+        return false;
+      }
+    }
+  }
+  
+  return true; // 安全
+}
+
+void GVNContext::invalidateMemoryValues(StoreInst* store) {
+  unsigned storePtr = getValueNumber(store->getPointer());
+  
+  if (DEBUG) {
+    std::cout << "        Invalidating memory values affected by store" << std::endl;
+  }
+  
+  // 找到所有可能被这个store影响的load表达式
+  std::vector<std::string> toRemove;
+  
+  for (auto& [exprKey, number] : expressionToNumber) {
+    if (exprKey.find("load_" + std::to_string(storePtr)) == 0) {
+      toRemove.push_back(exprKey);
+      if (DEBUG) {
+        std::cout << "          Invalidating expression: " << exprKey << std::endl;
+      }
+    }
+  }
+  
+  // 移除失效的表达式
+  for (const auto& key : toRemove) {
+    expressionToNumber.erase(key);
+  }
+}
+
+void GVNContext::eliminateRedundantInstructions(bool& changed) {
+  int removeCount = 0;
+  for (auto inst : needRemove) {
+    if (DEBUG) {
+      std::cout << "    Removing redundant instruction " << ++removeCount 
+                << "/" << needRemove.size() << ": " << inst->getName() << std::endl;
+    }
+    
+    // 删除指令前先断开所有使用关系
+    // inst->replaceAllUsesWith 已在 processInstruction 中调用
+    SysYIROptUtils::usedelete(inst);
+    changed = true;
+  }
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/GlobalStrengthReduction.cpp

@@ -0,0 +1,897 @@
+#include "GlobalStrengthReduction.h"
+#include "SysYIROptUtils.h"
+#include "IRBuilder.h"
+#include <algorithm>
+#include <cassert>
+#include <iostream>
+#include <cmath>
+
+extern int DEBUG;
+
+namespace sysy {
+
+// 全局强度削弱优化遍的静态 ID
+void *GlobalStrengthReduction::ID = (void *)&GlobalStrengthReduction::ID;
+
+// ======================================================================
+// GlobalStrengthReduction 类的实现
+// ======================================================================
+
+bool GlobalStrengthReduction::runOnFunction(Function *func, AnalysisManager &AM) {
+  if (func->getBasicBlocks().empty()) {
+    return false;
+  }
+
+  if (DEBUG) {
+    std::cout << "\n=== Running GlobalStrengthReduction on function: " << func->getName() << " ===" << std::endl;
+  }
+
+  bool changed = false;
+  GlobalStrengthReductionContext context(builder);
+  context.run(func, &AM, changed);
+
+  if (DEBUG) {
+    if (changed) {
+      std::cout << "GlobalStrengthReduction: Function " << func->getName() << " was modified" << std::endl;
+    } else {
+      std::cout << "GlobalStrengthReduction: Function " << func->getName() << " was not modified" << std::endl;
+    }
+    std::cout << "=== GlobalStrengthReduction completed for function: " << func->getName() << " ===" << std::endl;
+  }
+
+  return changed;
+}
+
+void GlobalStrengthReduction::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const {
+  // 强度削弱依赖副作用分析来判断指令是否可以安全优化
+  analysisDependencies.insert(&SysYSideEffectAnalysisPass::ID);
+  
+  // 强度削弱不会使分析失效，因为：
+  // - 只替换计算指令，不改变控制流
+  // - 不修改内存，不影响别名分析
+  // - 保持程序语义不变
+  // analysisInvalidations 保持为空
+  
+  if (DEBUG) {
+    std::cout << "GlobalStrengthReduction: Declared analysis dependencies (SideEffectAnalysis)" << std::endl;
+  }
+}
+
+// ======================================================================
+// GlobalStrengthReductionContext 类的实现
+// ======================================================================
+
+void GlobalStrengthReductionContext::run(Function *func, AnalysisManager *AM, bool &changed) {
+  if (DEBUG) {
+    std::cout << "  Starting GlobalStrengthReduction analysis for function: " << func->getName() << std::endl;
+  }
+
+  // 获取分析结果
+  if (AM) {
+    sideEffectAnalysis = AM->getAnalysisResult<SideEffectAnalysisResult, SysYSideEffectAnalysisPass>();
+    
+    if (DEBUG) {
+      if (sideEffectAnalysis) {
+        std::cout << "    GlobalStrengthReduction: Using side effect analysis" << std::endl;
+      } else {
+        std::cout << "    GlobalStrengthReduction: Warning - side effect analysis not available" << std::endl;
+      }
+    }
+  }
+
+  // 重置计数器
+  algebraicOptCount = 0;
+  strengthReductionCount = 0;
+  divisionOptCount = 0;
+
+  // 遍历所有基本块进行优化
+  for (auto &bb_ptr : func->getBasicBlocks()) {
+    if (processBasicBlock(bb_ptr.get())) {
+      changed = true;
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "  GlobalStrengthReduction completed for function: " << func->getName() << std::endl;
+    std::cout << "    Algebraic optimizations: " << algebraicOptCount << std::endl;
+    std::cout << "    Strength reductions: " << strengthReductionCount << std::endl;
+    std::cout << "    Division optimizations: " << divisionOptCount << std::endl;
+  }
+}
+
+bool GlobalStrengthReductionContext::processBasicBlock(BasicBlock *bb) {
+  bool changed = false;
+  
+  if (DEBUG) {
+    std::cout << "    Processing block: " << bb->getName() << std::endl;
+  }
+
+  // 收集需要处理的指令（避免迭代器失效）
+  std::vector<Instruction*> instructions;
+  for (auto &inst_ptr : bb->getInstructions()) {
+    instructions.push_back(inst_ptr.get());
+  }
+
+  // 处理每条指令
+  for (auto inst : instructions) {
+    if (processInstruction(inst)) {
+      changed = true;
+    }
+  }
+
+  return changed;
+}
+
+bool GlobalStrengthReductionContext::processInstruction(Instruction *inst) {
+
+  if (DEBUG) {
+    std::cout << "      Processing instruction: " << inst->getName() << std::endl;
+  }
+
+  // 先尝试代数优化
+  if (tryAlgebraicOptimization(inst)) {
+    algebraicOptCount++;
+    return true;
+  }
+
+  // 再尝试强度削弱
+  if (tryStrengthReduction(inst)) {
+    strengthReductionCount++;
+    return true;
+  }
+
+  return false;
+}
+
+// ======================================================================
+// 代数优化方法
+// ======================================================================
+
+bool GlobalStrengthReductionContext::tryAlgebraicOptimization(Instruction *inst) {
+  auto binary = dynamic_cast<BinaryInst*>(inst);
+  if (!binary) {
+    return false;
+  }
+
+  switch (binary->getKind()) {
+    case Instruction::kAdd:
+      return optimizeAddition(binary);
+    case Instruction::kSub:
+      return optimizeSubtraction(binary);
+    case Instruction::kMul:
+      return optimizeMultiplication(binary);
+    case Instruction::kDiv:
+      return optimizeDivision(binary);
+    case Instruction::kICmpEQ:
+    case Instruction::kICmpNE:
+    case Instruction::kICmpLT:
+    case Instruction::kICmpGT:
+    case Instruction::kICmpLE:
+    case Instruction::kICmpGE:
+      return optimizeComparison(binary);
+    case Instruction::kAnd:
+    case Instruction::kOr:
+      return optimizeLogical(binary);
+    default:
+      return false;
+  }
+}
+
+bool GlobalStrengthReductionContext::optimizeAddition(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+  int constVal;
+
+  // x + 0 = x
+  if (isConstantInt(rhs, constVal) && constVal == 0) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x + 0 -> x" << std::endl;
+    }
+    replaceWithOptimized(inst, lhs);
+    return true;
+  }
+  
+  // 0 + x = x
+  if (isConstantInt(lhs, constVal) && constVal == 0) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = 0 + x -> x" << std::endl;
+    }
+    replaceWithOptimized(inst, rhs);
+    return true;
+  }
+
+  // x + (-y) = x - y
+  if (auto rhsInst = dynamic_cast<UnaryInst*>(rhs)) {
+    if (rhsInst->getKind() == Instruction::kNeg) {
+      if (DEBUG) {
+        std::cout << "        Algebraic: " << inst->getName() << " = x + (-y) -> x - y" << std::endl;
+      }
+      // 创建减法指令
+      builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+      auto subInst = builder->createSubInst(lhs, rhsInst->getOperand());
+      replaceWithOptimized(inst, subInst);
+      return true;
+    }
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::optimizeSubtraction(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+  int constVal;
+
+  // x - 0 = x
+  if (isConstantInt(rhs, constVal) && constVal == 0) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x - 0 -> x" << std::endl;
+    }
+    replaceWithOptimized(inst, lhs);
+    return true;
+  }
+
+  // x - x = 0 (如果x没有副作用)
+  if (lhs == rhs && hasOnlyLocalUses(dynamic_cast<Instruction*>(lhs))) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x - x -> 0" << std::endl;
+    }
+    replaceWithOptimized(inst, getConstantInt(0));
+    return true;
+  }
+
+  // x - (-y) = x + y
+  if (auto rhsInst = dynamic_cast<UnaryInst*>(rhs)) {
+    if (rhsInst->getKind() == Instruction::kNeg) {
+      if (DEBUG) {
+        std::cout << "        Algebraic: " << inst->getName() << " = x - (-y) -> x + y" << std::endl;
+      }
+      builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+      auto addInst = builder->createAddInst(lhs, rhsInst->getOperand());
+      replaceWithOptimized(inst, addInst);
+      return true;
+    }
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::optimizeMultiplication(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+  int constVal;
+
+  // x * 0 = 0
+  if (isConstantInt(rhs, constVal) && constVal == 0) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x * 0 -> 0" << std::endl;
+    }
+    replaceWithOptimized(inst, getConstantInt(0));
+    return true;
+  }
+  
+  // 0 * x = 0
+  if (isConstantInt(lhs, constVal) && constVal == 0) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = 0 * x -> 0" << std::endl;
+    }
+    replaceWithOptimized(inst, getConstantInt(0));
+    return true;
+  }
+
+  // x * 1 = x
+  if (isConstantInt(rhs, constVal) && constVal == 1) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x * 1 -> x" << std::endl;
+    }
+    replaceWithOptimized(inst, lhs);
+    return true;
+  }
+  
+  // 1 * x = x
+  if (isConstantInt(lhs, constVal) && constVal == 1) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = 1 * x -> x" << std::endl;
+    }
+    replaceWithOptimized(inst, rhs);
+    return true;
+  }
+
+  // x * (-1) = -x
+  if (isConstantInt(rhs, constVal) && constVal == -1) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x * (-1) -> -x" << std::endl;
+    }
+    builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+    auto negInst = builder->createNegInst(lhs);
+    replaceWithOptimized(inst, negInst);
+    return true;
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::optimizeDivision(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+  int constVal;
+
+  // x / 1 = x
+  if (isConstantInt(rhs, constVal) && constVal == 1) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x / 1 -> x" << std::endl;
+    }
+    replaceWithOptimized(inst, lhs);
+    return true;
+  }
+
+  // x / (-1) = -x
+  if (isConstantInt(rhs, constVal) && constVal == -1) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x / (-1) -> -x" << std::endl;
+    }
+    builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+    auto negInst = builder->createNegInst(lhs);
+    replaceWithOptimized(inst, negInst);
+    return true;
+  }
+
+  // x / x = 1 (如果x != 0且没有副作用)
+  if (lhs == rhs && hasOnlyLocalUses(dynamic_cast<Instruction*>(lhs))) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x / x -> 1" << std::endl;
+    }
+    replaceWithOptimized(inst, getConstantInt(1));
+    return true;
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::optimizeComparison(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+
+  // x == x = true (如果x没有副作用)
+  if (inst->getKind() == Instruction::kICmpEQ && lhs == rhs && 
+      hasOnlyLocalUses(dynamic_cast<Instruction*>(lhs))) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x == x -> true" << std::endl;
+    }
+    replaceWithOptimized(inst, getConstantInt(1));
+    return true;
+  }
+
+  // x != x = false (如果x没有副作用)
+  if (inst->getKind() == Instruction::kICmpNE && lhs == rhs && 
+      hasOnlyLocalUses(dynamic_cast<Instruction*>(lhs))) {
+    if (DEBUG) {
+      std::cout << "        Algebraic: " << inst->getName() << " = x != x -> false" << std::endl;
+    }
+    replaceWithOptimized(inst, getConstantInt(0));
+    return true;
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::optimizeLogical(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+  int constVal;
+
+  if (inst->getKind() == Instruction::kAnd) {
+    // x && 0 = 0
+    if (isConstantInt(rhs, constVal) && constVal == 0) {
+      if (DEBUG) {
+        std::cout << "        Algebraic: " << inst->getName() << " = x && 0 -> 0" << std::endl;
+      }
+      replaceWithOptimized(inst, getConstantInt(0));
+      return true;
+    }
+    
+    // x && -1 = x
+    if (isConstantInt(rhs, constVal) && constVal == -1) {
+      if (DEBUG) {
+        std::cout << "        Algebraic: " << inst->getName() << " = x && 1 -> x" << std::endl;
+      }
+      replaceWithOptimized(inst, lhs);
+      return true;
+    }
+
+    // x && x = x
+    if (lhs == rhs) {
+      if (DEBUG) {
+        std::cout << "        Algebraic: " << inst->getName() << " = x && x -> x" << std::endl;
+      }
+      replaceWithOptimized(inst, lhs);
+      return true;
+    }
+  } else if (inst->getKind() == Instruction::kOr) {
+    // x || 0 = x
+    if (isConstantInt(rhs, constVal) && constVal == 0) {
+      if (DEBUG) {
+        std::cout << "        Algebraic: " << inst->getName() << " = x || 0 -> x" << std::endl;
+      }
+      replaceWithOptimized(inst, lhs);
+      return true;
+    }
+
+    // x || x = x
+    if (lhs == rhs) {
+      if (DEBUG) {
+        std::cout << "        Algebraic: " << inst->getName() << " = x || x -> x" << std::endl;
+      }
+      replaceWithOptimized(inst, lhs);
+      return true;
+    }
+  }
+
+  return false;
+}
+
+// ======================================================================
+// 强度削弱方法
+// ======================================================================
+
+bool GlobalStrengthReductionContext::tryStrengthReduction(Instruction *inst) {
+  if (auto binary = dynamic_cast<BinaryInst*>(inst)) {
+    switch (binary->getKind()) {
+      case Instruction::kMul:
+        return reduceMultiplication(binary);
+      case Instruction::kDiv:
+        return reduceDivision(binary);
+      default:
+        return false;
+    }
+  } else if (auto call = dynamic_cast<CallInst*>(inst)) {
+    return reducePower(call);
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::reduceMultiplication(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+  int constVal;
+
+  // 尝试右操作数为常数
+  Value* variable = lhs;
+  if (isConstantInt(rhs, constVal) && constVal > 0) {
+    return tryComplexMultiplication(inst, variable, constVal);
+  }
+  
+  // 尝试左操作数为常数
+  if (isConstantInt(lhs, constVal) && constVal > 0) {
+    variable = rhs;
+    return tryComplexMultiplication(inst, variable, constVal);
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::tryComplexMultiplication(BinaryInst* inst, Value* variable, int constant) {
+  // 首先检查是否为2的幂，使用简单位移
+  if (isPowerOfTwo(constant)) {
+    int shiftAmount = log2OfPowerOfTwo(constant);
+    if (DEBUG) {
+      std::cout << "        StrengthReduction: " << inst->getName() 
+                << " = x * " << constant << " -> x << " << shiftAmount << std::endl;
+    }
+    
+    builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+    auto shiftInst = builder->createBinaryInst(Instruction::kSll, Type::getIntType(), variable, getConstantInt(shiftAmount));
+    replaceWithOptimized(inst, shiftInst);
+    return true;
+  }
+  
+  // 尝试分解为位移和加法的组合
+  std::vector<int> shifts;
+  if (findOptimalShiftDecomposition(constant, shifts)) {
+    if (DEBUG) {
+      std::cout << "        StrengthReduction: " << inst->getName() 
+                << " = x * " << constant << " -> shift decomposition with " << shifts.size() << " terms" << std::endl;
+    }
+    
+    Value* result = createShiftDecomposition(inst, variable, shifts);
+    if (result) {
+      replaceWithOptimized(inst, result);
+      return true;
+    }
+  }
+  
+  return false;
+}
+
+bool GlobalStrengthReductionContext::findOptimalShiftDecomposition(int constant, std::vector<int>& shifts) {
+  shifts.clear();
+  
+  // 常见的有效分解模式
+  switch (constant) {
+    case 3:   // 3 = 2^1 + 2^0 -> (x << 1) + x
+      shifts = {1, 0};
+      return true;
+    case 5:   // 5 = 2^2 + 2^0 -> (x << 2) + x  
+      shifts = {2, 0};
+      return true;
+    case 6:   // 6 = 2^2 + 2^1 -> (x << 2) + (x << 1)
+      shifts = {2, 1};
+      return true;
+    case 7:   // 7 = 2^2 + 2^1 + 2^0 -> (x << 2) + (x << 1) + x
+      shifts = {2, 1, 0};
+      return true;
+    case 9:   // 9 = 2^3 + 2^0 -> (x << 3) + x
+      shifts = {3, 0};
+      return true;
+    case 10:  // 10 = 2^3 + 2^1 -> (x << 3) + (x << 1)
+      shifts = {3, 1};
+      return true;
+    case 11:  // 11 = 2^3 + 2^1 + 2^0 -> (x << 3) + (x << 1) + x
+      shifts = {3, 1, 0};
+      return true;
+    case 12:  // 12 = 2^3 + 2^2 -> (x << 3) + (x << 2)
+      shifts = {3, 2};
+      return true;
+    case 13:  // 13 = 2^3 + 2^2 + 2^0 -> (x << 3) + (x << 2) + x
+      shifts = {3, 2, 0};
+      return true;
+    case 14:  // 14 = 2^3 + 2^2 + 2^1 -> (x << 3) + (x << 2) + (x << 1)
+      shifts = {3, 2, 1};
+      return true;
+    case 15:  // 15 = 2^3 + 2^2 + 2^1 + 2^0 -> (x << 3) + (x << 2) + (x << 1) + x
+      shifts = {3, 2, 1, 0};
+      return true;
+    case 17:  // 17 = 2^4 + 2^0 -> (x << 4) + x
+      shifts = {4, 0};
+      return true;
+    case 18:  // 18 = 2^4 + 2^1 -> (x << 4) + (x << 1)
+      shifts = {4, 1};
+      return true;
+    case 20:  // 20 = 2^4 + 2^2 -> (x << 4) + (x << 2)
+      shifts = {4, 2};
+      return true;
+    case 24:  // 24 = 2^4 + 2^3 -> (x << 4) + (x << 3)
+      shifts = {4, 3};
+      return true;
+    case 25:  // 25 = 2^4 + 2^3 + 2^0 -> (x << 4) + (x << 3) + x
+      shifts = {4, 3, 0};
+      return true;
+    case 100: // 100 = 2^6 + 2^5 + 2^2 -> (x << 6) + (x << 5) + (x << 2)
+      shifts = {6, 5, 2};
+      return true;
+  }
+  
+  // 通用二进制分解（最多4个项，避免过度复杂化）
+  if (constant > 0 && constant < 256) {
+    std::vector<int> binaryShifts;
+    int temp = constant;
+    int bit = 0;
+    
+    while (temp > 0 && binaryShifts.size() < 4) {
+      if (temp & 1) {
+        binaryShifts.push_back(bit);
+      }
+      temp >>= 1;
+      bit++;
+    }
+    
+    // 只有当项数不超过3个时才使用二进制分解（比直接乘法更有效）
+    if (binaryShifts.size() <= 3 && binaryShifts.size() >= 2) {
+      shifts = binaryShifts;
+      return true;
+    }
+  }
+  
+  return false;
+}
+
+Value* GlobalStrengthReductionContext::createShiftDecomposition(BinaryInst* inst, Value* variable, const std::vector<int>& shifts) {
+  if (shifts.empty()) return nullptr;
+  
+  builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+  
+  Value* result = nullptr;
+  
+  for (int shift : shifts) {
+    Value* term;
+    if (shift == 0) {
+      // 0位移就是原变量
+      term = variable;
+    } else {
+      // 创建位移指令
+      term = builder->createBinaryInst(Instruction::kSll, Type::getIntType(), variable, getConstantInt(shift));
+    }
+    
+    if (result == nullptr) {
+      result = term;
+    } else {
+      // 累加到结果中
+      result = builder->createAddInst(result, term);
+    }
+  }
+  
+  return result;
+}
+
+bool GlobalStrengthReductionContext::reduceDivision(BinaryInst *inst) {
+  Value *lhs = inst->getLhs();
+  Value *rhs = inst->getRhs();
+  uint32_t constVal;
+
+  // x / 2^n = x >> n (对于无符号除法或已知为正数的情况)
+  if (isConstantInt(rhs, constVal) && constVal > 0 && isPowerOfTwo(constVal)) {
+    builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+    int shiftAmount = log2OfPowerOfTwo(constVal);
+    // 有符号除法校正：(x + (x >> 31) & mask) >> k
+    int maskValue = constVal - 1;
+    
+    // x >> 31 (算术右移获取符号位)
+    Value* signShift = ConstantInteger::get(31);
+    Value* signBits = builder->createBinaryInst(
+      Instruction::Kind::kSra,  // 算术右移
+      lhs->getType(),
+      lhs,
+      signShift
+    );
+    
+    // (x >> 31) & mask
+    Value* mask = ConstantInteger::get(maskValue);
+    Value* correction = builder->createBinaryInst(
+      Instruction::Kind::kAnd,
+      lhs->getType(),
+      signBits,
+      mask
+    );
+    
+    // x + correction
+    Value* corrected = builder->createAddInst(lhs, correction);
+    
+    // (x + correction) >> k
+    Value* divShift = ConstantInteger::get(shiftAmount);
+    Value* shiftInst = builder->createBinaryInst(
+      Instruction::Kind::kSra,  // 算术右移
+      lhs->getType(),
+      corrected,
+      divShift
+    );
+
+    if (DEBUG) {
+      std::cout << "        StrengthReduction: " << inst->getName() 
+                << " = x / " << constVal << " -> (x + (x >> 31) & mask) >> " << shiftAmount << std::endl;
+    }
+    
+    // builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+    // Value* divisor_minus_1 = ConstantInteger::get(constVal - 1);
+    // Value* adjusted = builder->createAddInst(lhs, divisor_minus_1);
+    // Value* shiftInst = builder->createBinaryInst(Instruction::kSra, Type::getIntType(), adjusted, getConstantInt(shiftAmount));
+    replaceWithOptimized(inst, shiftInst);
+    strengthReductionCount++;
+    return true;
+  }
+
+  // x / c = x * magic_number (魔数乘法优化 - 使用libdivide算法)
+  if (isConstantInt(rhs, constVal) && constVal > 1 && constVal != (uint32_t)(-1)) {
+    // auto magicPair = computeMulhMagicNumbers(static_cast<int>(constVal));
+    Value* magicResult = createMagicDivisionLibdivide(inst, static_cast<int>(constVal));
+    replaceWithOptimized(inst, magicResult);
+    divisionOptCount++;
+    return true;
+  }
+
+  return false;
+}
+
+bool GlobalStrengthReductionContext::reducePower(CallInst *inst) {
+  // 检查是否是pow函数调用
+  Function* callee = inst->getCallee();
+  if (!callee || callee->getName() != "pow") {
+    return false;
+  }
+
+  // pow(x, 2) = x * x
+  if (inst->getNumOperands() >= 2) {
+    int exponent;
+    if (isConstantInt(inst->getOperand(1), exponent)) {
+      if (exponent == 2) {
+        if (DEBUG) {
+          std::cout << "        StrengthReduction: pow(x, 2) -> x * x" << std::endl;
+        }
+        
+        Value* base = inst->getOperand(0);
+        builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+        auto mulInst = builder->createMulInst(base, base);
+        replaceWithOptimized(inst, mulInst);
+        strengthReductionCount++;
+        return true;
+      } else if (exponent >= 3 && exponent <= 8) {
+        // 对于小的指数，展开为连续乘法
+        if (DEBUG) {
+          std::cout << "        StrengthReduction: pow(x, " << exponent << ") -> repeated multiplication" << std::endl;
+        }
+        
+        Value* base = inst->getOperand(0);
+        Value* result = base;
+        builder->setPosition(inst->getParent(), inst->getParent()->findInstIterator(inst));
+        
+        for (int i = 1; i < exponent; i++) {
+          result = builder->createMulInst(result, base);
+        }
+        
+        replaceWithOptimized(inst, result);
+        strengthReductionCount++;
+        return true;
+      }
+    }
+  }
+
+  return false;
+}
+
+Value* GlobalStrengthReductionContext::createMagicDivisionLibdivide(BinaryInst* divInst, int divisor) {
+  builder->setPosition(divInst->getParent(), divInst->getParent()->findInstIterator(divInst));
+  // 使用mulh指令优化任意常数除法
+  auto [magic, shift] = SysYIROptUtils::computeMulhMagicNumbers(divisor);
+  
+  // 检查是否无法优化（magic == -1, shift == -1 表示失败）
+  if (magic == -1 && shift == -1) {
+    if (DEBUG) {
+      std::cout << "[SR] Cannot optimize division by " << divisor 
+                << ", keeping original division" << std::endl;
+    }
+    // 返回 nullptr 表示无法优化，调用方应该保持原始除法
+    return nullptr;
+  }
+  
+  // 2的幂次方除法可以用移位优化（但这不是魔数法的情况）这种情况应该不会被分类到这里但是还是做一个保护措施
+  if ((divisor & (divisor - 1)) == 0 && divisor > 0) {
+    // 是2的幂次方，可以用移位
+    int shift_amount = 0;
+    int temp = divisor;
+    while (temp > 1) {
+      temp >>= 1;
+      shift_amount++;
+    }
+    
+    Value* shiftConstant = ConstantInteger::get(shift_amount);
+    // 对于有符号除法，需要先加上除数-1然后再移位（为了正确处理负数舍入）
+    Value* divisor_minus_1 = ConstantInteger::get(divisor - 1);
+    Value* adjusted = builder->createAddInst(divInst->getOperand(0), divisor_minus_1);
+    return builder->createBinaryInst(
+      Instruction::Kind::kSra,  // 算术右移
+      divInst->getOperand(0)->getType(),
+      adjusted,
+      shiftConstant
+    );
+  }
+  
+  // 创建魔数常量
+  // 检查魔数是否能放入32位，如果不能，则不进行优化
+  if (magic > INT32_MAX || magic < INT32_MIN) {
+    if (DEBUG) {
+      std::cout << "[SR] Magic number " << magic << " exceeds 32-bit range, skipping optimization" << std::endl;
+    }
+    return nullptr; // 无法优化，保持原始除法
+  }
+  
+  Value* magicConstant = ConstantInteger::get((int32_t)magic);
+  
+  // 检查是否需要ADD_MARKER处理（加法调整）
+  bool needAdd = (shift & 0x40) != 0;
+  int actualShift = shift & 0x3F; // 提取真实的移位量
+  
+  if (DEBUG) {
+    std::cout << "[SR] IR Generation: magic=" << magic << ", needAdd=" << needAdd 
+              << ", actualShift=" << actualShift << std::endl;
+  }
+  
+  // 执行高位乘法：mulh(x, magic)
+  Value* mulhResult = builder->createBinaryInst(
+    Instruction::Kind::kMulh,  // 高位乘法
+    divInst->getOperand(0)->getType(),
+    divInst->getOperand(0),
+    magicConstant
+  );
+  
+  if (needAdd) {
+    // ADD_MARKER 情况：需要在移位前加上被除数
+    // 这对应于 libdivide 的加法调整算法
+    if (DEBUG) {
+      std::cout << "[SR] Applying ADD_MARKER: adding dividend before shift" << std::endl;
+    }
+    mulhResult = builder->createAddInst(mulhResult, divInst->getOperand(0));
+  }
+  
+  if (actualShift > 0) {
+    // 如果需要额外移位
+    Value* shiftConstant = ConstantInteger::get(actualShift);
+    mulhResult = builder->createBinaryInst(
+      Instruction::Kind::kSra,  // 算术右移
+      divInst->getOperand(0)->getType(),
+      mulhResult,
+      shiftConstant
+    );
+  }
+  
+  // 标准的有符号除法符号修正：如果被除数为负，商需要加1
+  // 这对所有有符号除法都需要，不管是否可能有负数
+  Value* isNegative = builder->createICmpLTInst(divInst->getOperand(0), ConstantInteger::get(0));
+  // 将i1转换为i32：负数时为1，非负数时为0 ICmpLTInst的结果会默认转化为32位
+  mulhResult = builder->createAddInst(mulhResult, isNegative);
+  
+  return mulhResult; 
+}
+
+// ======================================================================
+// 辅助方法
+// ======================================================================
+
+bool GlobalStrengthReductionContext::isPowerOfTwo(uint32_t n) {
+  return n > 0 && (n & (n - 1)) == 0;
+}
+
+int GlobalStrengthReductionContext::log2OfPowerOfTwo(uint32_t n) {
+  int result = 0;
+  while (n > 1) {
+    n >>= 1;
+    result++;
+  }
+  return result;
+}
+
+bool GlobalStrengthReductionContext::isConstantInt(Value* val, int& constVal) {
+  if (auto constInt = dynamic_cast<ConstantInteger*>(val)) {
+    constVal = std::get<int>(constInt->getVal());
+    return true;
+  }
+  return false;
+}
+
+bool GlobalStrengthReductionContext::isConstantInt(Value* val, uint32_t& constVal) {
+  if (auto constInt = dynamic_cast<ConstantInteger*>(val)) {
+    int signedVal = std::get<int>(constInt->getVal());
+    if (signedVal >= 0) {
+      constVal = static_cast<uint32_t>(signedVal);
+      return true;
+    }
+  }
+  return false;
+}
+
+ConstantInteger* GlobalStrengthReductionContext::getConstantInt(int val) {
+  return ConstantInteger::get(val);
+}
+
+bool GlobalStrengthReductionContext::hasOnlyLocalUses(Instruction* inst) {
+  if (!inst) return true;
+  
+  // 简单检查：如果指令没有副作用，则认为是本地的
+  if (sideEffectAnalysis) {
+    auto sideEffect = sideEffectAnalysis->getInstructionSideEffect(inst);
+    return sideEffect.type == SideEffectType::NO_SIDE_EFFECT;
+  }
+  
+  // 没有副作用分析时，保守处理
+  return !inst->isCall() && !inst->isStore() && !inst->isLoad();
+}
+
+void GlobalStrengthReductionContext::replaceWithOptimized(Instruction* original, Value* replacement) {
+  if (DEBUG) {
+    std::cout << "          Replacing " << original->getName() 
+              << " with " << replacement->getName() << std::endl;
+  }
+  
+  original->replaceAllUsesWith(replacement);
+  
+  // 如果替换值是新创建的指令，确保它有合适的名字
+//   if (auto replInst = dynamic_cast<Instruction*>(replacement)) {
+//     if (replInst->getName().empty()) {
+//       replInst->setName(original->getName() + "_opt");
+//     }
+//   }
+  
+  // 删除原指令，让调用者处理
+  SysYIROptUtils::usedelete(original);
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/InductionVariableElimination.cpp

@@ -0,0 +1,916 @@
+#include "InductionVariableElimination.h"
+#include "LoopCharacteristics.h"
+#include "Loop.h"
+#include "Dom.h"
+#include "SideEffectAnalysis.h"
+#include "AliasAnalysis.h"
+#include "SysYIROptUtils.h"
+#include <iostream>
+#include <algorithm>
+
+// 使用全局调试开关
+extern int DEBUG;
+
+namespace sysy {
+
+// 定义 Pass 的唯一 ID
+void *InductionVariableElimination::ID = (void *)&InductionVariableElimination::ID;
+
+bool InductionVariableElimination::runOnFunction(Function* F, AnalysisManager& AM) {
+  if (F->getBasicBlocks().empty()) {
+    return false; // 空函数
+  }
+
+  if (DEBUG) {
+    std::cout << "Running InductionVariableElimination on function: " << F->getName() << std::endl;
+  }
+
+  // 创建优化上下文并运行
+  InductionVariableEliminationContext context;
+  bool modified = context.run(F, AM);
+
+  if (DEBUG) {
+    std::cout << "InductionVariableElimination " << (modified ? "modified" : "did not modify") 
+              << " function: " << F->getName() << std::endl;
+  }
+
+  return modified;
+}
+
+void InductionVariableElimination::getAnalysisUsage(std::set<void*>& analysisDependencies, 
+                                                   std::set<void*>& analysisInvalidations) const {
+  // 依赖的分析
+  analysisDependencies.insert(&LoopAnalysisPass::ID);
+  analysisDependencies.insert(&LoopCharacteristicsPass::ID);
+  analysisDependencies.insert(&DominatorTreeAnalysisPass::ID);
+  analysisDependencies.insert(&SysYSideEffectAnalysisPass::ID);
+  analysisDependencies.insert(&SysYAliasAnalysisPass::ID);
+  
+  // 会使失效的分析（归纳变量消除会修改IR结构）
+  analysisInvalidations.insert(&LoopCharacteristicsPass::ID);
+  // 注意：支配树分析通常不会因为归纳变量消除而失效，因为我们不改变控制流
+}
+
+// ========== InductionVariableEliminationContext 实现 ==========
+
+bool InductionVariableEliminationContext::run(Function* F, AnalysisManager& AM) {
+  if (DEBUG) {
+    std::cout << "  Starting induction variable elimination analysis..." << std::endl;
+  }
+
+  // 获取必要的分析结果
+  loopAnalysis = AM.getAnalysisResult<LoopAnalysisResult, LoopAnalysisPass>(F);
+  if (!loopAnalysis || !loopAnalysis->hasLoops()) {
+    if (DEBUG) {
+      std::cout << "  No loops found, skipping induction variable elimination" << std::endl;
+    }
+    return false;
+  }
+
+  loopCharacteristics = AM.getAnalysisResult<LoopCharacteristicsResult, LoopCharacteristicsPass>(F);
+  if (!loopCharacteristics) {
+    if (DEBUG) {
+      std::cout << "  LoopCharacteristics analysis not available" << std::endl;
+    }
+    return false;
+  }
+
+  dominatorTree = AM.getAnalysisResult<DominatorTree, DominatorTreeAnalysisPass>(F);
+  if (!dominatorTree) {
+    if (DEBUG) {
+      std::cout << "  DominatorTree analysis not available" << std::endl;
+    }
+    return false;
+  }
+
+  sideEffectAnalysis = AM.getAnalysisResult<SideEffectAnalysisResult, SysYSideEffectAnalysisPass>();
+  if (!sideEffectAnalysis) {
+    if (DEBUG) {
+      std::cout << "  SideEffectAnalysis not available, using conservative approach" << std::endl;
+    }
+    // 可以继续执行，但会使用更保守的策略
+  } else {
+    if (DEBUG) {
+      std::cout << "  Using SideEffectAnalysis for safety checks" << std::endl;
+    }
+  }
+
+  aliasAnalysis = AM.getAnalysisResult<AliasAnalysisResult, SysYAliasAnalysisPass>(F);
+  if (!aliasAnalysis) {
+    if (DEBUG) {
+      std::cout << "  AliasAnalysis not available, using conservative approach" << std::endl;
+    }
+    // 可以继续执行，但会使用更保守的策略
+  } else {
+    if (DEBUG) {
+      std::cout << "  Using AliasAnalysis for memory safety checks" << std::endl;
+    }
+  }
+
+  // 执行三个阶段的优化
+  
+  // 阶段1：识别死归纳变量
+  identifyDeadInductionVariables(F);
+  
+  if (deadIVs.empty()) {
+    if (DEBUG) {
+      std::cout << "  No dead induction variables found" << std::endl;
+    }
+    return false;
+  }
+
+  if (DEBUG) {
+    std::cout << "  Found " << deadIVs.size() << " potentially dead induction variables" << std::endl;
+  }
+
+  // 阶段2：分析安全性
+  analyzeSafetyForElimination();
+
+  // 阶段3：执行消除
+  bool modified = performInductionVariableElimination();
+
+  if (DEBUG) {
+    printDebugInfo();
+  }
+
+  return modified;
+}
+
+void InductionVariableEliminationContext::identifyDeadInductionVariables(Function* F) {
+  if (DEBUG) {
+    std::cout << "  === Phase 1: Identifying Dead Induction Variables ===" << std::endl;
+  }
+
+  // 遍历所有循环
+  for (const auto& loop_ptr : loopAnalysis->getAllLoops()) {
+    Loop* loop = loop_ptr.get();
+    
+    if (DEBUG) {
+      std::cout << "    Analyzing loop: " << loop->getName() << std::endl;
+    }
+
+    // 获取循环特征
+    const LoopCharacteristics* characteristics = loopCharacteristics->getCharacteristics(loop);
+    if (!characteristics) {
+      if (DEBUG) {
+        std::cout << "      No characteristics available for loop" << std::endl;
+      }
+      continue;
+    }
+
+    if (characteristics->InductionVars.empty()) {
+      if (DEBUG) {
+        std::cout << "      No induction variables found in loop" << std::endl;
+      }
+      continue;
+    }
+
+    // 检查每个归纳变量是否为死归纳变量
+    for (const auto& iv : characteristics->InductionVars) {
+      auto deadIV = isDeadInductionVariable(iv.get(), loop);
+      if (deadIV) {
+        if (DEBUG) {
+          std::cout << "      Found potential dead IV: %" << deadIV->phiInst->getName() << std::endl;
+        }
+        
+        // 添加到候选项列表
+        loopToDeadIVs[loop].push_back(deadIV.get());
+        deadIVs.push_back(std::move(deadIV));
+      }
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "  === End Phase 1: Found " << deadIVs.size() << " candidates ===" << std::endl;
+  }
+}
+
+std::unique_ptr<DeadInductionVariable> 
+InductionVariableEliminationContext::isDeadInductionVariable(const InductionVarInfo* iv, Loop* loop) {
+  // 获取 phi 指令
+  auto* phiInst = dynamic_cast<PhiInst*>(iv->div);
+  if (!phiInst) {
+    return nullptr; // 不是 phi 指令
+  }
+
+  // 新的逻辑：递归分析整个use-def链，判断是否有真实的使用
+  if (!isPhiInstructionDeadRecursively(phiInst, loop)) {
+    return nullptr; // 有真实的使用，不能删除
+  }
+
+  // 创建死归纳变量信息
+  auto deadIV = std::make_unique<DeadInductionVariable>(phiInst, loop);
+  deadIV->relatedInsts = collectRelatedInstructions(phiInst, loop);
+  
+  return deadIV;
+}
+
+// 递归分析phi指令及其使用链是否都是死代码
+bool InductionVariableEliminationContext::isPhiInstructionDeadRecursively(PhiInst* phiInst, Loop* loop) {
+  if (DEBUG) {
+    std::cout << "      递归分析归纳变量 " << phiInst->getName() << " 的完整使用链" << std::endl;
+  }
+
+  // 使用访问集合避免无限递归
+  std::set<Instruction*> visitedInstructions;
+  std::set<Instruction*> currentPath; // 用于检测循环依赖
+  
+  // 核心逻辑：递归分析使用链，寻找任何"逃逸点"
+  return isInstructionUseChainDeadRecursively(phiInst, loop, visitedInstructions, currentPath);
+}
+
+// 递归分析指令的使用链是否都是死代码
+bool InductionVariableEliminationContext::isInstructionUseChainDeadRecursively(
+    Instruction* inst, Loop* loop, 
+    std::set<Instruction*>& visited, 
+    std::set<Instruction*>& currentPath) {
+  
+  if (DEBUG && visited.size() < 10) { // 限制debug输出
+    std::cout << "        分析指令 " << inst->getName() << " (" << inst->getKindString() << ")" << std::endl;
+  }
+  
+  // 避免无限递归
+  if (currentPath.count(inst) > 0) {
+    // 发现循环依赖，这在归纳变量中是正常的，继续分析其他路径
+    if (DEBUG && visited.size() < 10) {
+      std::cout << "          发现循环依赖，继续分析其他路径" << std::endl;
+    }
+    return true; // 循环依赖本身不是逃逸点
+  }
+  
+  if (visited.count(inst) > 0) {
+    // 已经分析过这个指令
+    return true; // 假设之前的分析是正确的
+  }
+  
+  visited.insert(inst);
+  currentPath.insert(inst);
+  
+  // 1. 检查是否有副作用（逃逸点）
+  if (sideEffectAnalysis && sideEffectAnalysis->hasSideEffect(inst)) {
+    if (DEBUG && visited.size() < 10) {
+      std::cout << "          指令有副作用，是逃逸点" << std::endl;
+    }
+    currentPath.erase(inst);
+    return false; // 有副作用的指令是逃逸点
+  }
+  
+  // 1.5. 特殊检查：控制流指令永远不是死代码
+  auto instKind = inst->getKind();
+  if (instKind == Instruction::Kind::kCondBr || 
+      instKind == Instruction::Kind::kBr ||
+      instKind == Instruction::Kind::kReturn) {
+    if (DEBUG && visited.size() < 10) {
+      std::cout << "          控制流指令，是逃逸点" << std::endl;
+    }
+    currentPath.erase(inst);
+    return false; // 控制流指令是逃逸点
+  }
+  
+  // 2. 检查指令的所有使用
+  bool allUsesAreDead = true;
+  for (auto use : inst->getUses()) {
+    auto user = use->getUser();
+    auto* userInst = dynamic_cast<Instruction*>(user);
+    
+    if (!userInst) {
+      // 被非指令使用（如函数返回值），是逃逸点
+      if (DEBUG && visited.size() < 10) {
+        std::cout << "          被非指令使用，是逃逸点" << std::endl;
+      }
+      allUsesAreDead = false;
+      break;
+    }
+    
+    // 检查使用是否在循环外（逃逸点）
+    if (!loop->contains(userInst->getParent())) {
+      if (DEBUG && visited.size() < 10) {
+        std::cout << "          在循环外被 " << userInst->getName() << " 使用，是逃逸点" << std::endl;
+      }
+      allUsesAreDead = false;
+      break;
+    }
+    
+    // 特殊检查：如果使用者是循环的退出条件，需要进一步分析
+    // 对于用于退出条件的归纳变量，需要更谨慎的处理
+    if (isUsedInLoopExitCondition(userInst, loop)) {
+      // 修复逻辑：用于循环退出条件的归纳变量通常不应该被消除
+      // 除非整个循环都可以被证明是完全无用的（这需要更复杂的分析）
+      if (DEBUG && visited.size() < 10) {
+        std::cout << "          被用于循环退出条件，是逃逸点（避免破坏循环语义）" << std::endl;
+      }
+      allUsesAreDead = false;
+      break;
+    }
+    
+    // 递归分析使用者的使用链
+    if (!isInstructionUseChainDeadRecursively(userInst, loop, visited, currentPath)) {
+      allUsesAreDead = false;
+      break; // 找到逃逸点，不需要继续分析
+    }
+  }
+  
+  currentPath.erase(inst);
+  
+  if (allUsesAreDead && DEBUG && visited.size() < 10) {
+    std::cout << "          指令 " << inst->getName() << " 的所有使用都是死代码" << std::endl;
+  }
+  
+  return allUsesAreDead;
+}
+
+// 检查循环是否有副作用
+bool InductionVariableEliminationContext::loopHasSideEffects(Loop* loop) {
+  // 遍历循环中的所有指令，检查是否有副作用
+  for (BasicBlock* bb : loop->getBlocks()) {
+    for (auto& inst : bb->getInstructions()) {
+      Instruction* instPtr = inst.get();
+      
+      // 使用副作用分析（如果可用）
+      if (sideEffectAnalysis && sideEffectAnalysis->hasSideEffect(instPtr)) {
+        if (DEBUG) {
+          std::cout << "          循环中发现有副作用的指令: " << instPtr->getName() << std::endl;
+        }
+        return true;
+      }
+      
+      // 如果没有副作用分析，使用保守的判断
+      if (!sideEffectAnalysis) {
+        auto kind = instPtr->getKind();
+        // 这些指令通常有副作用
+        if (kind == Instruction::Kind::kCall ||
+            kind == Instruction::Kind::kStore ||
+            kind == Instruction::Kind::kReturn) {
+          if (DEBUG) {
+            std::cout << "          循环中发现潜在有副作用的指令: " << instPtr->getName() << std::endl;
+          }
+          return true;
+        }
+      }
+    }
+  }
+  
+  // 重要修复：检查是否为嵌套循环的外层循环
+  // 如果当前循环包含其他循环，那么它有潜在的副作用
+  for (const auto& loop_ptr : loopAnalysis->getAllLoops()) {
+    Loop* otherLoop = loop_ptr.get();
+    if(loopAnalysis->getLowestCommonAncestor(otherLoop, loop) == loop) {
+      if (DEBUG) {
+        std::cout << "          循环 " << loop->getName() << " 是其他循环的外层循环，视为有副作用" << std::endl;
+      }
+      return true; // 外层循环被视为有副作用
+    }
+    // if (otherLoop != loop && loop->contains(otherLoop->getHeader())) {
+    //   if (DEBUG) {
+    //     std::cout << "          循环 " << loop->getName() << " 包含子循环 " << otherLoop->getName() << "，视为有副作用" << std::endl;
+    //   }
+    //   return true; // 包含子循环的外层循环被视为有副作用
+    // }
+  }
+  
+  if (DEBUG) {
+    std::cout << "          循环 " << loop->getName() << " 无副作用" << std::endl;
+  }
+  return false; // 循环无副作用
+}
+
+// 检查指令是否被用于循环退出条件
+bool InductionVariableEliminationContext::isUsedInLoopExitCondition(Instruction* inst, Loop* loop) {
+  // 检查指令是否被循环的退出条件使用
+  for (BasicBlock* exitingBB : loop->getExitingBlocks()) {
+    auto terminatorIt = exitingBB->terminator();
+    if (terminatorIt != exitingBB->end()) {
+      Instruction* terminator = terminatorIt->get();
+      if (terminator) {
+        // 检查终结指令的操作数
+        for (size_t i = 0; i < terminator->getNumOperands(); ++i) {
+          if (terminator->getOperand(i) == inst) {
+            if (DEBUG) {
+              std::cout << "          指令 " << inst->getName() << " 用于循环退出条件" << std::endl;
+            }
+            return true;
+          }
+        }
+        
+        // 对于条件分支，还需要检查条件指令的操作数
+        if (terminator->getKind() == Instruction::Kind::kCondBr) {
+          auto* condBr = dynamic_cast<CondBrInst*>(terminator);
+          if (condBr) {
+            Value* condition = condBr->getCondition();
+            if (condition == inst) {
+              if (DEBUG) {
+                std::cout << "          指令 " << inst->getName() << " 是循环条件" << std::endl;
+              }
+              return true;
+            }
+            
+            // 递归检查条件指令的操作数（比如比较指令）
+            auto* condInst = dynamic_cast<Instruction*>(condition);
+            if (condInst) {
+              for (size_t i = 0; i < condInst->getNumOperands(); ++i) {
+                if (condInst->getOperand(i) == inst) {
+                  if (DEBUG) {
+                    std::cout << "          指令 " << inst->getName() << " 用于循环条件的操作数" << std::endl;
+                  }
+                  return true;
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+  
+  return false;
+}
+
+
+// 检查指令的结果是否未被有效使用
+bool InductionVariableEliminationContext::isInstructionResultUnused(Instruction* inst, Loop* loop) {
+  // 检查指令的所有使用
+  if (inst->getUses().empty()) {
+    return true; // 没有使用，肯定是未使用
+  }
+  
+  for (auto use : inst->getUses()) {
+    auto user = use->getUser();
+    auto* userInst = dynamic_cast<Instruction*>(user);
+    
+    if (!userInst) {
+      return false; // 被非指令使用，认为是有效使用
+    }
+    
+    // 如果在循环外被使用，认为是有效使用
+    if (!loop->contains(userInst->getParent())) {
+      return false;
+    }
+    
+    // 递归检查使用这个结果的指令是否也是死代码
+    // 为了避免无限递归，限制递归深度
+    if (!isInstructionEffectivelyDead(userInst, loop, 3)) {
+      return false; // 存在有效使用
+    }
+  }
+  
+  return true; // 所有使用都是无效的
+}
+
+// 检查store指令是否存储到死地址（利用别名分析）
+bool InductionVariableEliminationContext::isStoreToDeadLocation(StoreInst* store, Loop* loop) {
+  if (!aliasAnalysis) {
+    return false; // 没有别名分析，保守返回false
+  }
+  
+  Value* storePtr = store->getPointer();
+  
+  // 检查是否存储到局部临时变量且该变量在循环外不被读取
+  const MemoryLocation* memLoc = aliasAnalysis->getMemoryLocation(storePtr);
+  if (!memLoc) {
+    return false; // 无法确定内存位置
+  }
+  
+  // 如果是局部数组且只在循环内被访问
+  if (memLoc->isLocalArray) {
+    // 检查该内存位置是否在循环外被读取
+    for (auto* accessInst : memLoc->accessInsts) {
+      if (accessInst->getKind() == Instruction::Kind::kLoad) {
+        if (!loop->contains(accessInst->getParent())) {
+          return false; // 在循环外被读取，不是死存储
+        }
+      }
+    }
+    
+    if (DEBUG) {
+      std::cout << "            存储到局部数组且仅在循环内访问" << std::endl;
+    }
+    return true; // 存储到仅循环内访问的局部数组
+  }
+  
+  return false;
+}
+
+// 检查指令是否有效死代码（带递归深度限制）
+bool InductionVariableEliminationContext::isInstructionEffectivelyDead(Instruction* inst, Loop* loop, int maxDepth) {
+  if (maxDepth <= 0) {
+    return false; // 达到递归深度限制，保守返回false
+  }
+  
+  // 利用副作用分析
+  if (sideEffectAnalysis && sideEffectAnalysis->hasSideEffect(inst)) {
+    return false; // 有副作用的指令不是死代码
+  }
+  
+  // 检查特殊指令类型
+  switch (inst->getKind()) {
+    case Instruction::Kind::kStore:
+      // Store指令可能是死存储
+      return isStoreToDeadLocation(dynamic_cast<StoreInst*>(inst), loop);
+    
+    case Instruction::Kind::kCall:
+      // 函数调用通常有副作用
+      if (sideEffectAnalysis) {
+        return !sideEffectAnalysis->hasSideEffect(inst);
+      }
+      return false; // 保守地认为函数调用有效果
+    
+    case Instruction::Kind::kReturn:
+    case Instruction::Kind::kBr:
+    case Instruction::Kind::kCondBr:
+      // 控制流指令不是死代码
+      return false;
+    
+    default:
+      // 其他指令检查其使用是否有效
+      break;
+  }
+  
+  // 检查指令的使用
+  if (inst->getUses().empty()) {
+    return true; // 没有使用的纯指令是死代码
+  }
+  
+  // 递归检查所有使用
+  for (auto use : inst->getUses()) {
+    auto user = use->getUser();
+    auto* userInst = dynamic_cast<Instruction*>(user);
+    
+    if (!userInst) {
+      return false; // 被非指令使用
+    }
+    
+    if (!loop->contains(userInst->getParent())) {
+      return false; // 在循环外被使用
+    }
+    
+    // 递归检查使用者
+    if (!isInstructionEffectivelyDead(userInst, loop, maxDepth - 1)) {
+      return false; // 存在有效使用
+    }
+  }
+  
+  return true; // 所有使用都是死代码
+}
+
+// 原有的函数保持兼容，但现在使用增强的死代码分析
+bool InductionVariableEliminationContext::isInstructionDeadOrInternalOnly(Instruction* inst, Loop* loop) {
+  return isInstructionEffectivelyDead(inst, loop, 5);
+}
+
+// 检查store指令是否有后续的load操作
+bool InductionVariableEliminationContext::hasSubsequentLoad(StoreInst* store, Loop* loop) {
+  if (!aliasAnalysis) {
+    // 没有别名分析，保守地假设有后续读取
+    return true;
+  }
+  
+  Value* storePtr = store->getPointer();
+  const MemoryLocation* storeLoc = aliasAnalysis->getMemoryLocation(storePtr);
+  
+  if (!storeLoc) {
+    // 无法确定内存位置，保守处理
+    return true;
+  }
+  
+  // 在循环中和循环后查找对同一位置的load操作
+  std::vector<BasicBlock*> blocksToCheck;
+  
+  // 添加循环内的所有基本块
+  for (auto* bb : loop->getBlocks()) {
+    blocksToCheck.push_back(bb);
+  }
+  
+  // 添加循环的退出块
+  auto exitBlocks = loop->getExitBlocks();
+  for (auto* exitBB : exitBlocks) {
+    blocksToCheck.push_back(exitBB);
+  }
+  
+  // 搜索load操作
+  for (auto* bb : blocksToCheck) {
+    for (auto& inst : bb->getInstructions()) {
+      if (inst->getKind() == Instruction::Kind::kLoad) {
+        LoadInst* loadInst = static_cast<LoadInst*>(inst.get());
+        Value* loadPtr = loadInst->getPointer();
+        const MemoryLocation* loadLoc = aliasAnalysis->getMemoryLocation(loadPtr);
+        
+        if (loadLoc && aliasAnalysis->queryAlias(storePtr, loadPtr) != AliasType::NO_ALIAS) {
+          // 找到可能读取同一位置的load操作
+          if (DEBUG) {
+            std::cout << "            找到后续load操作: " << loadInst->getName() << std::endl;
+          }
+          return true;
+        }
+      }
+    }
+  }
+  
+  // 检查是否通过函数调用间接访问
+  for (auto* bb : blocksToCheck) {
+    for (auto& inst : bb->getInstructions()) {
+      if (inst->getKind() == Instruction::Kind::kCall) {
+        CallInst* callInst = static_cast<CallInst*>(inst.get());
+        if (callInst && sideEffectAnalysis && sideEffectAnalysis->hasSideEffect(callInst)) {
+          // 函数调用可能间接读取内存
+          if (DEBUG) {
+            std::cout << "            函数调用可能读取内存: " << callInst->getName() << std::endl;
+          }
+          return true;
+        }
+      }
+    }
+  }
+  
+  if (DEBUG) {
+    std::cout << "            未找到后续load操作" << std::endl;
+  }
+  return false; // 没有找到后续读取
+}
+
+// 检查指令是否在循环外有使用
+bool InductionVariableEliminationContext::hasUsageOutsideLoop(Instruction* inst, Loop* loop) {
+  for (auto use : inst->getUses()) {
+    auto user = use->getUser();
+    auto* userInst = dynamic_cast<Instruction*>(user);
+    
+    if (!userInst) {
+      // 被非指令使用，可能在循环外
+      return true;
+    }
+    
+    if (!loop->contains(userInst->getParent())) {
+      // 在循环外被使用
+      if (DEBUG) {
+        std::cout << "            指令 " << inst->getName() << " 在循环外被 " 
+                  << userInst->getName() << " 使用" << std::endl;
+      }
+      return true;
+    }
+  }
+  
+  return false; // 没有循环外使用
+}
+
+// 检查store指令是否在循环外有后续的load操作
+bool InductionVariableEliminationContext::hasSubsequentLoadOutsideLoop(StoreInst* store, Loop* loop) {
+  if (!aliasAnalysis) {
+    // 没有别名分析，保守地假设有后续读取
+    return true;
+  }
+  
+  Value* storePtr = store->getPointer();
+  
+  // 检查循环的退出块及其后继
+  auto exitBlocks = loop->getExitBlocks();
+  std::set<BasicBlock*> visitedBlocks;
+  
+  for (auto* exitBB : exitBlocks) {
+    if (hasLoadInSubtree(exitBB, storePtr, visitedBlocks)) {
+      if (DEBUG) {
+        std::cout << "            找到循环外的后续load操作" << std::endl;
+      }
+      return true;
+    }
+  }
+  
+  return false; // 没有找到循环外的后续读取
+}
+
+// 递归检查基本块子树中是否有对指定位置的load操作
+bool InductionVariableEliminationContext::hasLoadInSubtree(BasicBlock* bb, Value* ptr, std::set<BasicBlock*>& visited) {
+  if (visited.count(bb) > 0) {
+    return false; // 已经访问过，避免无限循环
+  }
+  visited.insert(bb);
+  
+  // 检查当前基本块中的指令
+  for (auto& inst : bb->getInstructions()) {
+    if (inst->getKind() == Instruction::Kind::kLoad) {
+      LoadInst* loadInst = static_cast<LoadInst*>(inst.get());
+      if (aliasAnalysis && aliasAnalysis->queryAlias(ptr, loadInst->getPointer()) != AliasType::NO_ALIAS) {
+        return true; // 找到了对相同或别名位置的load
+      }
+    } else if (inst->getKind() == Instruction::Kind::kCall) {
+      // 函数调用可能间接读取内存
+      CallInst* callInst = static_cast<CallInst*>(inst.get());
+      if (sideEffectAnalysis && sideEffectAnalysis->hasSideEffect(callInst)) {
+        return true; // 保守地认为函数调用可能读取内存
+      }
+    }
+  }
+  
+  // 递归检查后继基本块（限制深度以避免过度搜索）
+  static int searchDepth = 0;
+  if (searchDepth < 10) { // 限制搜索深度
+    searchDepth++;
+    for (auto* succ : bb->getSuccessors()) {
+      if (hasLoadInSubtree(succ, ptr, visited)) {
+        searchDepth--;
+        return true;
+      }
+    }
+    searchDepth--;
+  }
+  
+  return false;
+}
+
+std::vector<Instruction*> InductionVariableEliminationContext::collectRelatedInstructions(
+    PhiInst* phiInst, Loop* loop) {
+  std::vector<Instruction*> relatedInsts;
+  
+  // 收集所有与该归纳变量相关的指令
+  for (auto use : phiInst->getUses()) {
+    auto user = use->getUser();
+    auto* userInst = dynamic_cast<Instruction*>(user);
+    
+    if (userInst && loop->contains(userInst->getParent())) {
+      relatedInsts.push_back(userInst);
+    }
+  }
+  
+  return relatedInsts;
+}
+
+void InductionVariableEliminationContext::analyzeSafetyForElimination() {
+  if (DEBUG) {
+    std::cout << "  === Phase 2: Analyzing Safety for Elimination ===" << std::endl;
+  }
+
+  // 为每个死归纳变量检查消除的安全性
+  for (auto& deadIV : deadIVs) {
+    bool isSafe = isSafeToEliminate(deadIV.get());
+    deadIV->canEliminate = isSafe;
+    
+    if (DEBUG) {
+      std::cout << "    Dead IV " << deadIV->phiInst->getName() 
+                << ": " << (isSafe ? "SAFE" : "UNSAFE") << " to eliminate" << std::endl;
+    }
+  }
+
+  if (DEBUG) {
+    size_t safeCount = 0;
+    for (const auto& deadIV : deadIVs) {
+      if (deadIV->canEliminate) safeCount++;
+    }
+    std::cout << "  === End Phase 2: " << safeCount << " of " << deadIVs.size() 
+              << " variables are safe to eliminate ===" << std::endl;
+  }
+}
+
+bool InductionVariableEliminationContext::isSafeToEliminate(const DeadInductionVariable* deadIV) {
+  // 1. 确保归纳变量在循环头
+  if (deadIV->phiInst->getParent() != deadIV->containingLoop->getHeader()) {
+    if (DEBUG) {
+      std::cout << "      Unsafe: phi not in loop header" << std::endl;
+    }
+    return false;
+  }
+  
+  // 2. 确保相关指令都在循环内
+  for (auto* inst : deadIV->relatedInsts) {
+    if (!deadIV->containingLoop->contains(inst->getParent())) {
+      if (DEBUG) {
+        std::cout << "      Unsafe: related instruction outside loop" << std::endl;
+      }
+      return false;
+    }
+  }
+  
+  // 3. 确保没有副作用
+  for (auto* inst : deadIV->relatedInsts) {
+    if (sideEffectAnalysis) {
+      // 使用副作用分析进行精确检查
+      if (sideEffectAnalysis->hasSideEffect(inst)) {
+        if (DEBUG) {
+          std::cout << "      Unsafe: related instruction " << inst->getName() 
+                    << " has side effects" << std::endl;
+        }
+        return false;
+      }
+    } else {
+      // 没有副作用分析时使用保守策略：只允许基本算术运算
+      auto kind = inst->getKind();
+      if (kind != Instruction::Kind::kAdd && 
+          kind != Instruction::Kind::kSub &&
+          kind != Instruction::Kind::kMul) {
+        if (DEBUG) {
+          std::cout << "      Unsafe: related instruction may have side effects (conservative)" << std::endl;
+        }
+        return false;
+      }
+    }
+  }
+  
+  // 4. 确保不影响循环的退出条件
+  for (BasicBlock* exitingBB : deadIV->containingLoop->getExitingBlocks()) {
+    auto terminatorIt = exitingBB->terminator();
+    if (terminatorIt != exitingBB->end()) {
+      Instruction* terminator = terminatorIt->get();
+      if (terminator) {
+        for (size_t i = 0; i < terminator->getNumOperands(); ++i) {
+          if (terminator->getOperand(i) == deadIV->phiInst) {
+            if (DEBUG) {
+              std::cout << "      Unsafe: phi used in loop exit condition" << std::endl;
+            }
+            return false;
+          }
+        }
+      }
+    }
+  }
+  
+  return true;
+}
+
+bool InductionVariableEliminationContext::performInductionVariableElimination() {
+  if (DEBUG) {
+    std::cout << "  === Phase 3: Performing Induction Variable Elimination ===" << std::endl;
+  }
+
+  bool modified = false;
+
+  for (auto& deadIV : deadIVs) {
+    if (!deadIV->canEliminate) {
+      continue;
+    }
+
+    if (DEBUG) {
+      std::cout << "    Eliminating dead IV: " << deadIV->phiInst->getName() << std::endl;
+    }
+
+    if (eliminateDeadInductionVariable(deadIV.get())) {
+      if (DEBUG) {
+        std::cout << "      Successfully eliminated: " << deadIV->phiInst->getName() << std::endl;
+      }
+      modified = true;
+    } else {
+      if (DEBUG) {
+        std::cout << "      Failed to eliminate: " << deadIV->phiInst->getName() << std::endl;
+      }
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "  === End Phase 3: " << (modified ? "Eliminations performed" : "No eliminations") << " ===" << std::endl;
+  }
+
+  return modified;
+}
+
+bool InductionVariableEliminationContext::eliminateDeadInductionVariable(DeadInductionVariable* deadIV) {
+  // 1. 删除所有相关指令
+  for (auto* inst : deadIV->relatedInsts) {
+    auto* bb = inst->getParent();
+    auto it = bb->findInstIterator(inst);
+    if (it != bb->end()) {
+        SysYIROptUtils::usedelete(it);
+    //   bb->getInstructions().erase(it);
+      if (DEBUG) {
+        std::cout << "        Removed related instruction: " << inst->getName() << std::endl;
+      }
+    }
+  }
+
+  // 2. 删除 phi 指令
+  auto* bb = deadIV->phiInst->getParent();
+  auto it = bb->findInstIterator(deadIV->phiInst);
+  if (it != bb->end()) {
+    SysYIROptUtils::usedelete(it);
+    // bb->getInstructions().erase(it);
+    if (DEBUG) {
+      std::cout << "        Removed phi instruction: " << deadIV->phiInst->getName() << std::endl;
+    }
+    return true;
+  }
+
+  return false;
+}
+
+void InductionVariableEliminationContext::printDebugInfo() {
+  if (!DEBUG) return;
+
+  std::cout << "\n=== Induction Variable Elimination Summary ===" << std::endl;
+  std::cout << "Total dead IVs found: " << deadIVs.size() << std::endl;
+  
+  size_t eliminatedCount = 0;
+  for (auto& [loop, loopDeadIVs] : loopToDeadIVs) {
+    size_t loopEliminatedCount = 0;
+    for (auto* deadIV : loopDeadIVs) {
+      if (deadIV->canEliminate) {
+        loopEliminatedCount++;
+        eliminatedCount++;
+      }
+    }
+    
+    if (loopEliminatedCount > 0) {
+      std::cout << "Loop " << loop->getName() << ": " << loopEliminatedCount 
+                << " of " << loopDeadIVs.size() << " IVs eliminated" << std::endl;
+    }
+  }
+  
+  std::cout << "Total eliminated: " << eliminatedCount << " of " << deadIVs.size() << std::endl;
+  std::cout << "=============================================" << std::endl;
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/LICM.cpp

@@ -0,0 +1,264 @@
+#include "LICM.h"
+#include "IR.h"
+
+extern int DEBUG;
+
+namespace sysy {
+
+void *LICM::ID = (void *)&LICM::ID;
+
+bool LICMContext::run() { return hoistInstructions(); }
+
+bool LICMContext::hoistInstructions() {
+  bool changed = false;
+  BasicBlock *preheader = loop->getPreHeader();
+  if (!preheader || !chars)
+    return false;
+
+  // 1. 先收集所有可外提指令
+  std::unordered_set<Instruction *> workSet(chars->invariantInsts.begin(), chars->invariantInsts.end());
+
+  if (DEBUG) {
+    std::cout << "LICM: Found " << workSet.size() << " candidate invariant instructions to hoist:" << std::endl;
+    for (auto *inst : workSet) {
+      std::cout << "  - " << inst->getName() << " (kind: " << static_cast<int>(inst->getKind()) 
+                << ", in BB: " << inst->getParent()->getName() << ")" << std::endl;
+    }
+  }
+
+  // 2. 计算每个指令被依赖的次数（入度）
+  std::unordered_map<Instruction *, int> indegree;
+  std::unordered_map<Instruction *, std::vector<Instruction *>> dependencies; // 记录依赖关系
+  std::unordered_map<Instruction *, std::vector<Instruction *>> dependents;   // 记录被依赖关系
+  
+  for (auto *inst : workSet) {
+    indegree[inst] = 0;
+    dependencies[inst] = {};
+    dependents[inst] = {};
+  }
+  
+  if (DEBUG) {
+    std::cout << "LICM: Analyzing dependencies between invariant instructions..." << std::endl;
+  }
+  
+  for (auto *inst : workSet) {
+    for (size_t i = 0; i < inst->getNumOperands(); ++i) {
+      if (auto *dep = dynamic_cast<Instruction *>(inst->getOperand(i))) {
+        if (workSet.count(dep)) {
+          indegree[inst]++;
+          dependencies[inst].push_back(dep);
+          dependents[dep].push_back(inst);
+          
+          if (DEBUG) {
+            std::cout << "  Dependency: " << inst->getName() << " depends on " << dep->getName() << std::endl;
+          }
+        }
+      }
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "LICM: Initial indegree analysis:" << std::endl;
+    for (auto &[inst, deg] : indegree) {
+      std::cout << "  " << inst->getName() << ": indegree=" << deg;
+      if (deg > 0) {
+        std::cout << ", depends on: ";
+        for (auto *dep : dependencies[inst]) {
+          std::cout << dep->getName() << " ";
+        }
+      }
+      std::cout << std::endl;
+    }
+  }
+
+  // 3. Kahn拓扑排序
+  std::vector<Instruction *> sorted;
+  std::queue<Instruction *> q;
+  
+  if (DEBUG) {
+    std::cout << "LICM: Starting topological sort..." << std::endl;
+  }
+  
+  for (auto &[inst, deg] : indegree) {
+    if (deg == 0) {
+      q.push(inst);
+      if (DEBUG) {
+        std::cout << "  Initial zero-indegree instruction: " << inst->getName() << std::endl;
+      }
+    }
+  }
+  
+  int sortStep = 0;
+  while (!q.empty()) {
+    auto *inst = q.front();
+    q.pop();
+    sorted.push_back(inst);
+    
+    if (DEBUG) {
+      std::cout << "  Step " << (++sortStep) << ": Processing " << inst->getName() << std::endl;
+    }
+    
+    if (DEBUG) {
+      std::cout << "    Reducing indegree of dependents of " << inst->getName() << std::endl;
+    }
+    
+    // 正确的拓扑排序：当处理一个指令时，应该减少其所有使用者（dependents）的入度
+    for (auto *dependent : dependents[inst]) {
+      indegree[dependent]--;
+      if (DEBUG) {
+        std::cout << "      Reducing indegree of " << dependent->getName() << " to " << indegree[dependent] << std::endl;
+      }
+      if (indegree[dependent] == 0) {
+        q.push(dependent);
+        if (DEBUG) {
+          std::cout << "      Adding " << dependent->getName() << " to queue (indegree=0)" << std::endl;
+        }
+      }
+    }
+  }
+
+  // 检查是否全部排序，若未全部排序，打印错误信息
+  // 这可能是因为存在循环依赖或其他问题导致无法完成拓扑排序
+  if (sorted.size() != workSet.size()) {
+    if (DEBUG) {
+      std::cout << "LICM: Topological sort failed! Sorted " << sorted.size() 
+                << " instructions out of " << workSet.size() << " total." << std::endl;
+      
+      // 找出未被排序的指令（形成循环依赖的指令）
+      std::unordered_set<Instruction *> remaining;
+      for (auto *inst : workSet) {
+        bool found = false;
+        for (auto *sortedInst : sorted) {
+          if (inst == sortedInst) {
+            found = true;
+            break;
+          }
+        }
+        if (!found) {
+          remaining.insert(inst);
+        }
+      }
+      
+      std::cout << "LICM: Instructions involved in dependency cycle:" << std::endl;
+      for (auto *inst : remaining) {
+        std::cout << "  - " << inst->getName() << " (indegree=" << indegree[inst] << ")" << std::endl;
+        std::cout << "    Dependencies within cycle: ";
+        for (auto *dep : dependencies[inst]) {
+          if (remaining.count(dep)) {
+            std::cout << dep->getName() << " ";
+          }
+        }
+        std::cout << std::endl;
+        std::cout << "    Dependents within cycle: ";
+        for (auto *dependent : dependents[inst]) {
+          if (remaining.count(dependent)) {
+            std::cout << dependent->getName() << " ";
+          }
+        }
+        std::cout << std::endl;
+      }
+      
+      // 尝试找出一个具体的循环路径
+      std::cout << "LICM: Attempting to trace a dependency cycle:" << std::endl;
+      if (!remaining.empty()) {
+        auto *start = *remaining.begin();
+        std::unordered_set<Instruction *> visited;
+        std::vector<Instruction *> path;
+        
+        std::function<bool(Instruction *)> findCycle = [&](Instruction *current) -> bool {
+          if (visited.count(current)) {
+            // 找到环
+            auto it = std::find(path.begin(), path.end(), current);
+            if (it != path.end()) {
+              std::cout << "  Cycle found: ";
+              for (auto cycleIt = it; cycleIt != path.end(); ++cycleIt) {
+                std::cout << (*cycleIt)->getName() << " -> ";
+              }
+              std::cout << current->getName() << std::endl;
+              return true;
+            }
+            return false;
+          }
+          
+          visited.insert(current);
+          path.push_back(current);
+          
+          for (auto *dep : dependencies[current]) {
+            if (remaining.count(dep)) {
+              if (findCycle(dep)) {
+                return true;
+              }
+            }
+          }
+          
+          path.pop_back();
+          return false;
+        };
+        
+        findCycle(start);
+      }
+    }
+    return false;
+  }
+
+  // 4. 按拓扑序外提
+  if (DEBUG) {
+    std::cout << "LICM: Successfully completed topological sort. Hoisting instructions in order:" << std::endl;
+  }
+  
+  for (auto *inst : sorted) {
+    if (!inst)
+      continue;
+    BasicBlock *parent = inst->getParent();
+    if (parent && loop->contains(parent)) {
+      if (DEBUG) {
+        std::cout << "  Hoisting " << inst->getName() << " from " << parent->getName() 
+                  << " to preheader " << preheader->getName() << std::endl;
+      }
+      auto sourcePos = parent->findInstIterator(inst);
+      auto targetPos = preheader->terminator();
+      parent->moveInst(sourcePos, targetPos, preheader);
+      changed = true;
+    }
+  }
+  
+  if (DEBUG && changed) {
+    std::cout << "LICM: Successfully hoisted " << sorted.size() << " invariant instructions" << std::endl;
+  }
+  
+  return changed;
+}
+// ---- LICM Pass Implementation ----
+
+bool LICM::runOnFunction(Function *F, AnalysisManager &AM) {
+  auto *loopAnalysis = AM.getAnalysisResult<LoopAnalysisResult, LoopAnalysisPass>(F);
+  auto *loopCharsResult = AM.getAnalysisResult<LoopCharacteristicsResult, LoopCharacteristicsPass>(F);
+  if (!loopAnalysis || !loopCharsResult)
+    return false;
+
+  bool changed = false;
+  // 对每个函数内的所有循环做处理
+  for (const auto &loop_ptr : loopAnalysis->getAllLoops()) {
+    Loop *loop = loop_ptr.get();
+    if (DEBUG) {
+      std::cout << "LICM: Processing loop in function " << F->getName() << ": " << loop->getName() << std::endl;
+    }
+    const LoopCharacteristics *chars = loopCharsResult->getCharacteristics(loop);
+    if (!chars || !loop->getPreHeader())
+      continue; // 没有分析结果或没有前置块则跳过
+    LICMContext ctx(F, loop, builder, chars);
+    changed |= ctx.run();
+  }
+  return changed;
+}
+
+void LICM::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const {
+
+  analysisDependencies.insert(&LoopAnalysisPass::ID);
+  analysisDependencies.insert(&LoopCharacteristicsPass::ID);
+
+  analysisInvalidations.insert(&LoopCharacteristicsPass::ID);
+  analysisInvalidations.insert(&LivenessAnalysisPass::ID);
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/LargeArrayToGlobal.cpp

@@ -0,0 +1,145 @@
+#include "../../include/midend/Pass/Optimize/LargeArrayToGlobal.h"
+#include "../../IR.h"
+#include <unordered_map>
+#include <sstream>
+#include <string>
+
+namespace sysy {
+
+// Helper function to convert type to string
+static std::string typeToString(Type *type) {
+    if (!type) return "null";
+    
+    switch (type->getKind()) {
+        case Type::kInt:
+            return "int";
+        case Type::kFloat:
+            return "float";
+        case Type::kPointer:
+            return "ptr";
+        case Type::kArray: {
+            auto *arrayType = type->as<ArrayType>();
+            return "[" + std::to_string(arrayType->getNumElements()) + " x " + 
+                   typeToString(arrayType->getElementType()) + "]";
+        }
+        default:
+            return "unknown";
+    }
+}
+
+void *LargeArrayToGlobalPass::ID = &LargeArrayToGlobalPass::ID;
+
+bool LargeArrayToGlobalPass::runOnModule(Module *M, AnalysisManager &AM) {
+        bool changed = false;
+        
+        if (!M) {
+            return false;
+        }
+
+        // Collect all alloca instructions from all functions
+        std::vector<std::pair<AllocaInst*, Function*>> allocasToConvert;
+        
+        for (auto &funcPair : M->getFunctions()) {
+            Function *F = funcPair.second.get();
+            if (!F || F->getBasicBlocks().begin() == F->getBasicBlocks().end()) {
+                continue;
+            }
+            
+            for (auto &BB : F->getBasicBlocks()) {
+                for (auto &inst : BB->getInstructions()) {
+                    if (auto *alloca = dynamic_cast<AllocaInst*>(inst.get())) {
+                        Type *allocatedType = alloca->getAllocatedType();
+                        
+                        // Calculate the size of the allocated type
+                        unsigned size = calculateTypeSize(allocatedType);
+                        if(DEBUG){
+                            // Debug: print size information
+                             std::cout << "LargeArrayToGlobalPass: Found alloca with size " << size 
+                                  << " for type " << typeToString(allocatedType) << std::endl;
+                        }
+                        
+                        // Convert arrays of 1KB (1024 bytes) or larger to global variables
+                        if (size >= 1024) {
+                            if(DEBUG)
+                                std::cout << "LargeArrayToGlobalPass: Converting array of size " << size << " to global" << std::endl;
+                            allocasToConvert.emplace_back(alloca, F);
+                        }
+                    }
+                }
+            }
+        }
+
+        // Convert the collected alloca instructions to global variables
+        for (auto [alloca, F] : allocasToConvert) {
+            convertAllocaToGlobal(alloca, F, M);
+            changed = true;
+        }
+
+return changed;
+    }
+
+unsigned LargeArrayToGlobalPass::calculateTypeSize(Type *type) {
+    if (!type) return 0;
+
+    switch (type->getKind()) {
+        case Type::kInt:
+        case Type::kFloat:
+            return 4;
+        case Type::kPointer:
+            return 8;
+        case Type::kArray: {
+            auto *arrayType = type->as<ArrayType>();
+            return arrayType->getNumElements() * calculateTypeSize(arrayType->getElementType());
+        }
+        default:
+            return 0;
+    }
+}
+
+void LargeArrayToGlobalPass::convertAllocaToGlobal(AllocaInst *alloca, Function *F, Module *M) {
+    Type *allocatedType = alloca->getAllocatedType();
+    
+    // Create a unique name for the global variable
+    std::string globalName = generateUniqueGlobalName(alloca, F);
+    
+    // Create the global variable - GlobalValue expects pointer type
+    Type *pointerType = Type::getPointerType(allocatedType);
+    GlobalValue *globalVar = M->createGlobalValue(globalName, pointerType);
+    
+    if (!globalVar) {
+        return;
+    }
+    
+    // Replace all uses of the alloca with the global variable
+    alloca->replaceAllUsesWith(globalVar);
+    
+    // Remove the alloca instruction from its basic block
+    for (auto &BB : F->getBasicBlocks()) {
+        auto &instructions = BB->getInstructions();
+        for (auto it = instructions.begin(); it != instructions.end(); ++it) {
+            if (it->get() == alloca) {
+                instructions.erase(it);
+                break;
+            }
+        }
+    }
+}
+
+std::string LargeArrayToGlobalPass::generateUniqueGlobalName(AllocaInst *alloca, Function *F) {
+    std::string baseName = alloca->getName();
+    if (baseName.empty()) {
+        baseName = "array";
+    }
+    
+    // Ensure uniqueness by appending function name and counter
+    static std::unordered_map<std::string, int> nameCounter;
+    std::string key = F->getName() + "." + baseName;
+    
+    int counter = nameCounter[key]++;
+    std::ostringstream oss;
+    oss << key << "." << counter;
+    
+    return oss.str();
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/LoopNormalization.cpp

@@ -0,0 +1,528 @@
+#include "LoopNormalization.h"
+#include "Dom.h"
+#include "Loop.h"
+#include "SysYIROptUtils.h"
+#include <iostream>
+#include <algorithm>
+#include <sstream>
+
+// 使用全局调试开关
+extern int DEBUG;
+
+namespace sysy {
+
+// 定义 Pass 的唯一 ID
+void *LoopNormalizationPass::ID = (void *)&LoopNormalizationPass::ID;
+
+bool LoopNormalizationPass::runOnFunction(Function *F, AnalysisManager &AM) {
+  if (F->getBasicBlocks().empty()) {
+    return false; // 空函数
+  }
+
+  if (DEBUG)
+    std::cout << "Running LoopNormalizationPass on function: " << F->getName() << std::endl;
+
+  // 获取并缓存所有需要的分析结果
+  loopAnalysis = AM.getAnalysisResult<LoopAnalysisResult, LoopAnalysisPass>(F);
+  if (!loopAnalysis || !loopAnalysis->hasLoops()) {
+    if (DEBUG)
+      std::cout << "No loops found in function " << F->getName() << ", skipping normalization" << std::endl;
+    return false; // 没有循环需要规范化
+  }
+
+  domTree = AM.getAnalysisResult<DominatorTree, DominatorTreeAnalysisPass>(F);
+  
+  if (!domTree) {
+    std::cerr << "Error: DominatorTree not available for function " << F->getName() << std::endl;
+    return false;
+  }
+
+  // 重置统计信息
+  stats = NormalizationStats();
+  
+  bool modified = false;
+  const auto& allLoops = loopAnalysis->getAllLoops();
+  stats.totalLoops = allLoops.size();
+
+  if (DEBUG) {
+    std::cout << "Found " << stats.totalLoops << " loops to analyze for normalization" << std::endl;
+  }
+
+  // 按循环深度从外到内处理，确保外层循环先规范化
+  std::vector<Loop*> sortedLoops;
+  for (const auto& loop_ptr : allLoops) {
+    sortedLoops.push_back(loop_ptr.get());
+  }
+  
+  std::sort(sortedLoops.begin(), sortedLoops.end(), [](Loop* a, Loop* b) {
+    return a->getLoopDepth() < b->getLoopDepth(); // 按深度升序排列
+  });
+
+  // 逐个规范化循环
+  for (Loop* loop : sortedLoops) {
+    if (needsPreheader(loop)) {
+      stats.loopsNeedingPreheader++;
+      
+      if (DEBUG) {
+        std::cout << "  Loop " << loop->getName() << " needs preheader (depth=" 
+                  << loop->getLoopDepth() << ")" << std::endl;
+      }
+      
+      if (normalizeLoop(loop)) {
+        modified = true;
+        stats.loopsNormalized++;
+        
+        // 验证规范化结果
+        if (!validateNormalization(loop)) {
+          std::cerr << "Warning: Loop normalization validation failed for loop " 
+                    << loop->getName() << std::endl;
+        }
+      }
+    } else {
+      if (DEBUG) {
+        auto* preheader = getExistingPreheader(loop);
+        if (preheader) {
+          std::cout << "  Loop " << loop->getName() << " already has preheader: " 
+                    << preheader->getName() << std::endl;
+        }
+      }
+    }
+  }
+
+  if (DEBUG && modified) {
+    printStats(F);
+  }
+
+  return modified;
+}
+
+bool LoopNormalizationPass::normalizeLoop(Loop* loop) {
+  if (DEBUG)
+    std::cout << "    Normalizing loop: " << loop->getName() << std::endl;
+
+  // 创建前置块
+  BasicBlock* preheader = createPreheaderForLoop(loop);
+  if (!preheader) {
+    if (DEBUG)
+      std::cout << "    Failed to create preheader for loop " << loop->getName() << std::endl;
+    return false;
+  }
+
+  stats.preheadersCreated++;
+  
+  if (DEBUG) {
+    std::cout << "    Successfully created preheader " << preheader->getName() 
+              << " for loop " << loop->getName() << std::endl;
+  }
+
+  return true;
+}
+
+BasicBlock* LoopNormalizationPass::createPreheaderForLoop(Loop* loop) {
+  BasicBlock* header = loop->getHeader();
+  if (!header) {
+    if (DEBUG)
+      std::cerr << "    Error: Loop has no header block" << std::endl;
+    return nullptr;
+  }
+
+  // 获取循环外的前驱块
+  std::vector<BasicBlock*> externalPreds = getExternalPredecessors(loop);
+  if (externalPreds.empty()) {
+    if (DEBUG)
+      std::cout << "    Loop " << loop->getName() << " has no external predecessors" << std::endl;
+    return nullptr;
+  }
+
+  if (DEBUG) {
+    std::cout << "    Found " << externalPreds.size() << " external predecessors for loop " 
+              << loop->getName() << std::endl;
+    for (auto* pred : externalPreds) {
+      std::cout << "      External pred: " << pred->getName() << std::endl;
+    }
+  }
+
+  // 生成前置块名称
+  std::string preheaderName = generatePreheaderName(loop);
+  
+  // 创建新的前置块
+  Function* parentFunction = header->getParent();
+  BasicBlock* preheader = parentFunction->addBasicBlock(preheaderName, header);
+  
+  if (!preheader) {
+    if (DEBUG)
+      std::cerr << "    Error: Failed to create basic block " << preheaderName << std::endl;
+    return nullptr;
+  }
+
+  // 在前置块中创建跳转指令到循环头部
+  builder->setPosition(preheader, preheader->end());
+  UncondBrInst* br = builder->createUncondBrInst(header);
+
+  // 更新preheader的CFG关系
+  preheader->addSuccessor(header);
+  header->addPredecessor(preheader);
+
+  if(DEBUG) {
+    std::cout << "    Created preheader " << preheader->getName() 
+              << " with unconditional branch to " << header->getName() << std::endl;
+  }
+  // 重定向外部前驱到新的前置块
+  redirectExternalPredecessors(loop, preheader, header, externalPreds);
+
+  // 更新PHI节点
+  updatePhiNodesForPreheader(header, preheader, externalPreds);
+
+  // 更新支配树关系
+  updateDominatorRelations(preheader, loop);
+
+  // 重要：更新循环对象的前置块信息
+  // 这样后续的优化遍可以通过 loop->getPreHeader() 获取到新创建的前置块
+  loop->setPreHeader(preheader);
+
+  if (DEBUG) {
+    std::cout << "    Updated loop object: preheader set to " << preheader->getName() << std::endl;
+  }
+
+  return preheader;
+}
+
+bool LoopNormalizationPass::needsPreheader(Loop* loop) {
+  // 检查是否已有合适的前置块
+  if (getExistingPreheader(loop) != nullptr) {
+    return false;
+  }
+
+  // 检查是否有外部前驱（如果没有外部前驱，不需要前置块）
+  std::vector<BasicBlock*> externalPreds = getExternalPredecessors(loop);
+  if (externalPreds.empty()) {
+    return false;
+  }
+
+  // 基于结构性需求判断：
+  // 1. 如果有多个外部前驱，必须创建前置块来合并它们
+  // 2. 如果单个外部前驱不适合作为前置块，需要创建新的前置块
+  return (externalPreds.size() > 1) || !isSuitableAsPreheader(externalPreds[0], loop);
+}
+
+BasicBlock* LoopNormalizationPass::getExistingPreheader(Loop* loop) {
+  BasicBlock* header = loop->getHeader();
+  if (!header) return nullptr;
+
+  std::vector<BasicBlock*> externalPreds = getExternalPredecessors(loop);
+  
+  // 如果只有一个外部前驱，且适合作为前置块，则返回它
+  if (externalPreds.size() == 1 && isSuitableAsPreheader(externalPreds[0], loop)) {
+    return externalPreds[0];
+  }
+
+  return nullptr;
+}
+
+void LoopNormalizationPass::updateDominatorRelations(BasicBlock* newBlock, Loop* loop) {
+  // 由于在getAnalysisUsage中声明了DominatorTree会失效，
+  // PassManager会在本遍运行后自动将支配树结果标记为失效，
+  // 后续需要支配树的Pass会触发重新计算，所以这里无需手动更新
+  
+  if (DEBUG) {
+    BasicBlock* header = loop->getHeader();
+    std::cout << "    DominatorTree marked for invalidation - new preheader " 
+              << newBlock->getName() << " will dominate " << header->getName() 
+              << " after recomputation by PassManager" << std::endl;
+  }
+}
+
+void LoopNormalizationPass::redirectExternalPredecessors(Loop* loop, BasicBlock* preheader, BasicBlock* header, 
+                                                         const std::vector<BasicBlock*>& externalPreds) {
+  // std::vector<BasicBlock*> externalPreds = getExternalPredecessors(loop);
+  
+  if (DEBUG) {
+    std::cout << "    Redirecting " << externalPreds.size() << " external predecessors" << std::endl;
+  }
+
+  for (BasicBlock* pred : externalPreds) {
+    // 获取前驱块的终止指令
+    auto termIt = pred->terminator();
+    if (termIt == pred->end()) continue;
+    
+    Instruction* terminator = termIt->get();
+    if (!terminator) continue;
+
+    // 更新跳转目标
+    if (auto* br = dynamic_cast<UncondBrInst*>(terminator)) {
+      // 无条件跳转
+      if (br->getBlock() == header) {
+        if(DEBUG){
+          std::cout << "      Updating unconditional branch from " << br->getBlock()->getName()
+                    << " to " << preheader->getName() << std::endl;
+        }
+        // 需要更新操作数
+        br->setOperand(0, preheader);
+        // 更新CFG关系
+        header->removePredecessor(pred);
+        preheader->addPredecessor(pred);
+        pred->removeSuccessor(header);
+        pred->addSuccessor(preheader);
+        
+      }
+    } else if (auto* condBr = dynamic_cast<CondBrInst*>(terminator)) {
+      // 条件跳转
+      bool updated = false;
+      if (condBr->getThenBlock() == header) {
+        condBr->setOperand(1, preheader);  // 第1个操作数是then分支
+        updated = true;
+      }
+      if (condBr->getElseBlock() == header) {
+        condBr->setOperand(2, preheader);  // 第2个操作数是else分支
+        updated = true;
+      }
+      if (updated) {
+        // 更新CFG关系
+        header->removePredecessor(pred);
+        preheader->addPredecessor(pred);
+        pred->removeSuccessor(header);
+        pred->addSuccessor(preheader);
+        
+        if (DEBUG) {
+          std::cout << "      Updated conditional branch from " << pred->getName() 
+                    << " to " << preheader->getName() << std::endl;
+        }
+      }
+    }
+  }
+}
+
+std::string LoopNormalizationPass::generatePreheaderName(Loop* loop) {
+  std::ostringstream oss;
+  oss << loop->getName() << "_preheader";
+  return oss.str();
+}
+
+bool LoopNormalizationPass::validateNormalization(Loop* loop) {
+  BasicBlock* header = loop->getHeader();
+  if (!header) return false;
+
+  // 检查循环是否现在有唯一的外部前驱
+  std::vector<BasicBlock*> externalPreds = getExternalPredecessors(loop);
+  if (externalPreds.size() != 1) {
+    if (DEBUG)
+      std::cout << "    Validation failed: Loop " << loop->getName() 
+                << " has " << externalPreds.size() << " external predecessors (expected 1)" << std::endl;
+    return false;
+  }
+
+  // 检查外部前驱是否适合作为前置块
+  BasicBlock* preheader = externalPreds[0];
+  if (!isSuitableAsPreheader(preheader, loop)) {
+    if (DEBUG)
+      std::cout << "    Validation failed: External predecessor " << preheader->getName() 
+                << " is not suitable as preheader" << std::endl;
+    return false;
+  }
+
+  // 额外验证：检查CFG连接性
+  if (!preheader->hasSuccessor(header)) {
+    if (DEBUG)
+      std::cout << "    Validation failed: Preheader " << preheader->getName() 
+                << " is not connected to header " << header->getName() << std::endl;
+    return false;
+  }
+
+  if (!header->hasPredecessor(preheader)) {
+    if (DEBUG)
+      std::cout << "    Validation failed: Header " << header->getName() 
+                << " does not have preheader " << preheader->getName() << " as predecessor" << std::endl;
+    return false;
+  }
+
+  if (DEBUG)
+    std::cout << "    Validation passed for loop " << loop->getName() << std::endl;
+  
+  return true;
+}
+
+std::vector<BasicBlock*> LoopNormalizationPass::getExternalPredecessors(Loop* loop) {
+  std::vector<BasicBlock*> externalPreds;
+  BasicBlock* header = loop->getHeader();
+  if (!header) return externalPreds;
+
+  for (BasicBlock* pred : header->getPredecessors()) {
+    if (!loop->contains(pred)) {
+      externalPreds.push_back(pred);
+    }
+  }
+
+  return externalPreds;
+}
+
+bool LoopNormalizationPass::isSuitableAsPreheader(BasicBlock* block, Loop* loop) {
+  if (!block) return false;
+
+  // 检查该块是否只有一个后继，且后继是循环头部
+  auto successors = block->getSuccessors();
+  if (successors.size() != 1) {
+    return false;
+  }
+
+  if (successors[0] != loop->getHeader()) {
+    return false;
+  }
+
+  // 检查该块是否不包含复杂的控制流
+  // 理想的前置块应该只包含简单的跳转指令
+  size_t instCount = 0;
+  for (const auto& inst : block->getInstructions()) {
+    instCount++;
+    // 如果指令过多，可能不适合作为前置块
+    if (instCount > 10) { // 阈值可调整
+      return false;
+    }
+  }
+
+  return true;
+}
+
+void LoopNormalizationPass::updatePhiNodesForPreheader(BasicBlock* header, BasicBlock* preheader,
+                                                      const std::vector<BasicBlock*>& oldPreds) {
+  if (DEBUG) {
+    std::cout << "    Updating PHI nodes in header " << header->getName() 
+              << " for new preheader " << preheader->getName() << std::endl;
+  }
+
+  std::vector<PhiInst*> phisToRemove; // 需要删除的PHI节点
+  
+  for (auto& inst : header->getInstructions()) {
+    if (auto* phi = dynamic_cast<PhiInst*>(inst.get())) {
+      if (DEBUG) {
+        std::cout << "      Processing PHI node: " << phi->getName() << std::endl;
+      }
+
+      // 收集来自外部前驱的值 - 需要保持原始的映射关系
+      std::map<BasicBlock*, Value*> externalValues;
+      for (BasicBlock* oldPred : oldPreds) {
+        Value* value = phi->getValfromBlk(oldPred);
+        if (value) {
+          externalValues[oldPred] = value;
+        }
+      }
+
+      // 处理PHI节点的更新
+      if (externalValues.size() > 1) {
+        // 多个外部前驱：在前置块中创建新的PHI节点
+        builder->setPosition(preheader, preheader->getInstructions().begin());
+        
+        std::vector<Value*> values;
+        std::vector<BasicBlock*> blocks;
+        for (auto& [block, value] : externalValues) {
+          values.push_back(value);
+          blocks.push_back(block);
+        }
+        
+        PhiInst* newPhi = builder->createPhiInst(phi->getType(), values, blocks);
+        
+        // 移除所有外部前驱的条目
+        for (BasicBlock* oldPred : oldPreds) {
+          phi->removeIncomingBlock(oldPred);
+        }
+        
+        // 添加来自新前置块的条目
+        phi->addIncoming(newPhi, preheader);
+        
+      } else if (externalValues.size() == 1) {
+        // 单个外部前驱：直接重新映射
+        Value* value = externalValues.begin()->second;
+        
+        // 移除旧的外部前驱条目
+        for (BasicBlock* oldPred : oldPreds) {
+          phi->removeIncomingBlock(oldPred);
+        }
+        
+        // 添加来自新前置块的条目
+        phi->addIncoming(value, preheader);
+        
+        // 检查PHI节点是否只剩下一个条目（只来自前置块）
+        if (phi->getNumIncomingValues() == 1) {
+          if (DEBUG) {
+            std::cout << "      PHI node " << phi->getName() 
+                      << " now has only one incoming value, scheduling for removal" << std::endl;
+          }
+          // 用单一值替换所有使用
+          Value* singleValue = phi->getIncomingValue(0u);
+          phi->replaceAllUsesWith(singleValue);
+          phisToRemove.push_back(phi);
+        }
+      } else {
+        // 没有外部值的PHI节点：检查是否需要更新
+        // 这种PHI节点只有循环内的边，通常不需要修改
+        // 但我们仍然需要检查是否只有一个条目
+        if (phi->getNumIncomingValues() == 1) {
+          if (DEBUG) {
+            std::cout << "      PHI node " << phi->getName() 
+                      << " has only one incoming value (no external), scheduling for removal" << std::endl;
+          }
+          // 用单一值替换所有使用
+          Value* singleValue = phi->getIncomingValue(0u);
+          phi->replaceAllUsesWith(singleValue);
+          phisToRemove.push_back(phi);
+        }
+      }
+      
+      if (DEBUG && std::find(phisToRemove.begin(), phisToRemove.end(), phi) == phisToRemove.end()) {
+        std::cout << "      Updated PHI node with " << externalValues.size() 
+                  << " external values, total incoming: " << phi->getNumIncomingValues() << std::endl;
+      }
+    }
+  }
+  
+  // 删除标记为移除的PHI节点
+  for (PhiInst* phi : phisToRemove) {
+    if (DEBUG) {
+      std::cout << "      Removing redundant PHI node: " << phi->getName() << std::endl;
+    }
+    SysYIROptUtils::usedelete(phi);
+  }
+  
+  // 更新统计信息
+  stats.redundantPhisRemoved += phisToRemove.size();
+  
+  if (DEBUG && !phisToRemove.empty()) {
+    std::cout << "    Removed " << phisToRemove.size() << " redundant PHI nodes" << std::endl;
+  }
+}
+
+void LoopNormalizationPass::printStats(Function* F) {
+  std::cout << "\n--- Loop Normalization Statistics for Function: " << F->getName() << " ---" << std::endl;
+  std::cout << "Total loops analyzed: " << stats.totalLoops << std::endl;
+  std::cout << "Loops needing preheader: " << stats.loopsNeedingPreheader << std::endl;
+  std::cout << "Preheaders created: " << stats.preheadersCreated << std::endl;
+  std::cout << "Loops successfully normalized: " << stats.loopsNormalized << std::endl;
+  std::cout << "Redundant PHI nodes removed: " << stats.redundantPhisRemoved << std::endl;
+  
+  if (stats.totalLoops > 0) {
+    double normalizationRate = (double)stats.loopsNormalized / stats.totalLoops * 100.0;
+    std::cout << "Normalization rate: " << normalizationRate << "%" << std::endl;
+  }
+  
+  std::cout << "---------------------------------------------------------------" << std::endl;
+}
+
+void LoopNormalizationPass::getAnalysisUsage(std::set<void *> &analysisDependencies, 
+                                            std::set<void *> &analysisInvalidations) const {
+  // LoopNormalization依赖的分析
+  analysisDependencies.insert(&LoopAnalysisPass::ID);              // 循环结构分析
+  analysisDependencies.insert(&DominatorTreeAnalysisPass::ID);     // 支配树分析
+
+  // LoopNormalization会修改CFG结构，因此会使以下分析失效
+  analysisInvalidations.insert(&DominatorTreeAnalysisPass::ID);    // 支配树需要重新计算
+  
+  // 注意：我们不让循环结构分析失效，原因如下：
+  // 1. 循环规范化只添加前置块，不改变循环的核心结构（头部、体、回边）
+  // 2. 我们会手动更新Loop对象的前置块信息（通过loop->setPreHeader()）
+  // 3. 让循环分析失效并重新计算的成本较高且不必要
+  // 4. 后续优化遍可以正确获取到更新后的前置块信息
+  // 
+  // 如果未来有更复杂的循环结构修改，可能需要考虑让循环分析失效：
+  // analysisInvalidations.insert(&LoopAnalysisPass::ID);
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/LoopStrengthReduction.cpp

@@ -0,0 +1,940 @@
+#include "LoopStrengthReduction.h"
+#include "LoopCharacteristics.h"
+#include "Loop.h"
+#include "Dom.h"
+#include "IRBuilder.h"
+#include "SysYIROptUtils.h"
+#include <iostream>
+#include <algorithm>
+#include <cmath>
+#include <unordered_map>
+#include <climits>
+
+// 使用全局调试开关
+extern int DEBUG;
+
+namespace sysy {
+
+// 定义 Pass 
+void *LoopStrengthReduction::ID = (void *)&LoopStrengthReduction::ID;
+
+bool StrengthReductionContext::analyzeInductionVariableRange(
+  const InductionVarInfo* ivInfo, 
+  Loop* loop
+) const {
+  if (!ivInfo->valid) {
+    if (DEBUG) {
+      std::cout << "        Invalid IV info, assuming potential negative" << std::endl;
+    }
+    return true; // 保守假设非线性变化可能为负数
+  }
+
+  // 获取phi指令的所有入口值
+  auto* phiInst = dynamic_cast<PhiInst*>(ivInfo->base);
+  if (!phiInst) {
+    if (DEBUG) {
+      std::cout << "        No phi instruction, assuming potential negative" << std::endl;
+    }
+    return true; // 无法确定，保守假设
+  }
+
+  bool hasNegativePotential = false;
+  bool hasNonNegativeInitial = false;
+  int initialValue = 0;
+  
+  for (auto& [incomingBB, incomingVal] : phiInst->getIncomingValues()) {
+    // 检查初始值（来自循环外的值）
+    if (!loop->contains(incomingBB)) {
+      if (auto* constInt = dynamic_cast<ConstantInteger*>(incomingVal)) {
+        initialValue = constInt->getInt();
+        if (initialValue < 0) {
+          if (DEBUG) {
+            std::cout << "        Found negative initial value: " << initialValue << std::endl;
+          }
+          hasNegativePotential = true;
+        } else {
+          if (DEBUG) {
+            std::cout << "        Found non-negative initial value: " << initialValue << std::endl;
+          }
+          hasNonNegativeInitial = true;
+        }
+      } else {
+        // 如果不是常数初始值，保守假设可能为负数
+        if (DEBUG) {
+          std::cout << "        Non-constant initial value, assuming potential negative" << std::endl;
+        }
+        hasNegativePotential = true;
+      }
+    }
+  }
+
+  // 检查递增值和偏移
+  if (ivInfo->factor < 0) {
+    if (DEBUG) {
+      std::cout << "        Negative factor: " << ivInfo->factor << std::endl;
+    }
+    hasNegativePotential = true;
+  }
+
+  if (ivInfo->offset < 0) {
+    if (DEBUG) {
+      std::cout << "        Negative offset: " << ivInfo->offset << std::endl;
+    }
+    hasNegativePotential = true;
+  }
+
+  // 精确分析：如果初始值非负，递增为正，偏移非负，则整个序列非负
+  if (hasNonNegativeInitial && ivInfo->factor > 0 && ivInfo->offset >= 0) {
+    if (DEBUG) {
+      std::cout << "        ANALYSIS: Confirmed non-negative range" << std::endl;
+      std::cout << "          Initial: " << initialValue << " >= 0" << std::endl;
+      std::cout << "          Factor: " << ivInfo->factor << " > 0" << std::endl;
+      std::cout << "          Offset: " << ivInfo->offset << " >= 0" << std::endl;
+    }
+    return false; // 确定不会为负数
+  }
+
+  // 报告分析结果
+  if (DEBUG) {
+    if (hasNegativePotential) {
+      std::cout << "        ANALYSIS: Potential negative values detected" << std::endl;
+    } else {
+      std::cout << "        ANALYSIS: No negative indicators, but missing positive confirmation" << std::endl;
+    }
+  }
+
+  return hasNegativePotential;
+}
+
+
+bool LoopStrengthReduction::runOnFunction(Function* F, AnalysisManager& AM) {
+  if (F->getBasicBlocks().empty()) {
+    return false; // 空函数
+  }
+
+  if (DEBUG) {
+    std::cout << "Running LoopStrengthReduction on function: " << F->getName() << std::endl;
+  }
+
+  // 创建优化上下文并运行
+  StrengthReductionContext context(builder);
+  bool modified = context.run(F, AM);
+
+  if (DEBUG) {
+    std::cout << "LoopStrengthReduction " << (modified ? "modified" : "did not modify") 
+              << " function: " << F->getName() << std::endl;
+  }
+
+  return modified;
+}
+
+void LoopStrengthReduction::getAnalysisUsage(std::set<void*>& analysisDependencies, 
+                                           std::set<void*>& analysisInvalidations) const {
+  // 依赖的分析
+  analysisDependencies.insert(&LoopAnalysisPass::ID);
+  analysisDependencies.insert(&LoopCharacteristicsPass::ID);
+  analysisDependencies.insert(&DominatorTreeAnalysisPass::ID);
+  
+  // 会使失效的分析（强度削弱会修改IR结构）
+  analysisInvalidations.insert(&LoopCharacteristicsPass::ID);
+  // 注意：支配树分析通常不会因为强度削弱而失效，因为我们不改变控制流
+}
+
+// ========== StrengthReductionContext 实现 ==========
+
+bool StrengthReductionContext::run(Function* F, AnalysisManager& AM) {
+  if (DEBUG) {
+    std::cout << "  Starting strength reduction analysis..." << std::endl;
+  }
+
+  // 获取必要的分析结果
+  loopAnalysis = AM.getAnalysisResult<LoopAnalysisResult, LoopAnalysisPass>(F);
+  if (!loopAnalysis || !loopAnalysis->hasLoops()) {
+    if (DEBUG) {
+      std::cout << "  No loops found, skipping strength reduction" << std::endl;
+    }
+    return false;
+  }
+
+  loopCharacteristics = AM.getAnalysisResult<LoopCharacteristicsResult, LoopCharacteristicsPass>(F);
+  if (!loopCharacteristics) {
+    if (DEBUG) {
+      std::cout << "  LoopCharacteristics analysis not available" << std::endl;
+    }
+    return false;
+  }
+
+  dominatorTree = AM.getAnalysisResult<DominatorTree, DominatorTreeAnalysisPass>(F);
+  if (!dominatorTree) {
+    if (DEBUG) {
+      std::cout << "  DominatorTree analysis not available" << std::endl;
+    }
+    return false;
+  }
+
+  // 执行三个阶段的优化
+  
+  // 阶段1：识别候选项
+  identifyStrengthReductionCandidates(F);
+  
+  if (candidates.empty()) {
+    if (DEBUG) {
+      std::cout << "  No strength reduction candidates found" << std::endl;
+    }
+    return false;
+  }
+
+  if (DEBUG) {
+    std::cout << "  Found " << candidates.size() << " potential candidates" << std::endl;
+  }
+
+  // 阶段2：分析优化潜力
+  analyzeOptimizationPotential();
+
+  // 阶段3：执行优化
+  bool modified = performStrengthReduction();
+
+  if (DEBUG) {
+    printDebugInfo();
+  }
+
+  return modified;
+}
+
+void StrengthReductionContext::identifyStrengthReductionCandidates(Function* F) {
+  if (DEBUG) {
+    std::cout << "  === Phase 1: Identifying Strength Reduction Candidates ===" << std::endl;
+  }
+
+  // 遍历所有循环
+  for (const auto& loop_ptr : loopAnalysis->getAllLoops()) {
+    Loop* loop = loop_ptr.get();
+    
+    if (DEBUG) {
+      std::cout << "    Analyzing loop: " << loop->getName() << std::endl;
+    }
+
+    // 获取循环特征
+    const LoopCharacteristics* characteristics = loopCharacteristics->getCharacteristics(loop);
+    if (!characteristics) {
+      if (DEBUG) {
+        std::cout << "      No characteristics available for loop" << std::endl;
+      }
+      continue;
+    }
+
+    if (characteristics->InductionVars.empty()) {
+      if (DEBUG) {
+        std::cout << "      No induction variables found in loop" << std::endl;
+      }
+      continue;
+    }
+
+    // 遍历循环中的所有指令
+    for (BasicBlock* bb : loop->getBlocks()) {
+      for (auto& inst_ptr : bb->getInstructions()) {
+        Instruction* inst = inst_ptr.get();
+        
+        // 检查是否为强度削弱候选项
+        auto candidate = isStrengthReductionCandidate(inst, loop);
+        if (candidate) {
+          if (DEBUG) {
+            std::cout << "      Found candidate: %" << inst->getName() 
+                      << " (IV: %" << candidate->inductionVar->getName() 
+                      << ", multiplier: " << candidate->multiplier 
+                      << ", offset: " << candidate->offset << ")" << std::endl;
+          }
+          
+          // 添加到候选项列表
+          loopToCandidates[loop].push_back(candidate.get());
+          candidates.push_back(std::move(candidate));
+        }
+      }
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "  === End Phase 1: Found " << candidates.size() << " candidates ===" << std::endl;
+  }
+}
+
+std::unique_ptr<StrengthReductionCandidate> 
+StrengthReductionContext::isStrengthReductionCandidate(Instruction* inst, Loop* loop) {
+  auto kind = inst->getKind();
+  
+  // 支持乘法、除法、取模指令
+  if (kind != Instruction::Kind::kMul && 
+      kind != Instruction::Kind::kDiv && 
+      kind != Instruction::Kind::kRem) {
+    return nullptr;
+  }
+
+  auto* binaryInst = dynamic_cast<BinaryInst*>(inst);
+  if (!binaryInst) {
+    return nullptr;
+  }
+
+  Value* op0 = binaryInst->getOperand(0);
+  Value* op1 = binaryInst->getOperand(1);
+
+  // 检查模式：归纳变量 op 常数 或 常数 op 归纳变量
+  Value* inductionVar = nullptr;
+  int constantValue = 0;
+  StrengthReductionCandidate::OpType opType;
+  
+  // 获取循环特征信息
+  const LoopCharacteristics* characteristics = loopCharacteristics->getCharacteristics(loop);
+  if (!characteristics) {
+    return nullptr;
+  }
+
+  // 确定操作类型
+  switch (kind) {
+    case Instruction::Kind::kMul:
+      opType = StrengthReductionCandidate::MULTIPLY;
+      break;
+    case Instruction::Kind::kDiv:
+      opType = StrengthReductionCandidate::DIVIDE;
+      break;
+    case Instruction::Kind::kRem:
+      opType = StrengthReductionCandidate::REMAINDER;
+      break;
+    default:
+      return nullptr;
+  }
+
+  // 模式1: IV op const
+  const InductionVarInfo* ivInfo = getInductionVarInfo(op0, loop, characteristics);
+  if (ivInfo && dynamic_cast<ConstantInteger*>(op1)) {
+    inductionVar = op0;
+    constantValue = dynamic_cast<ConstantInteger*>(op1)->getInt();
+  }
+  // 模式2: const op IV (仅对乘法有效)
+  else if (opType == StrengthReductionCandidate::MULTIPLY) {
+    ivInfo = getInductionVarInfo(op1, loop, characteristics);
+    if (ivInfo && dynamic_cast<ConstantInteger*>(op0)) {
+      inductionVar = op1;
+      constantValue = dynamic_cast<ConstantInteger*>(op0)->getInt();
+    }
+  }
+
+  if (!inductionVar || constantValue <= 1) {
+    return nullptr; // 不是有效的候选项
+  }
+
+  // 创建候选项
+  auto candidate = std::make_unique<StrengthReductionCandidate>(
+    inst, inductionVar, opType, constantValue, 0, inst->getParent(), loop
+  );
+
+  // 分析归纳变量是否可能为负数
+  candidate->hasNegativeValues = analyzeInductionVariableRange(ivInfo, loop);
+
+  // 根据除法类型选择优化策略
+  if (opType == StrengthReductionCandidate::DIVIDE) {
+    bool isPowerOfTwo = (constantValue & (constantValue - 1)) == 0;
+    
+    if (isPowerOfTwo) {
+      // 2的幂除法
+      if (candidate->hasNegativeValues) {
+        candidate->divStrategy = StrengthReductionCandidate::SIGNED_CORRECTION;
+        if (DEBUG) {
+          std::cout << "        Division by power of 2 with potential negative values, using signed correction" << std::endl;
+        }
+      } else {
+        candidate->divStrategy = StrengthReductionCandidate::SIMPLE_SHIFT;
+        if (DEBUG) {
+          std::cout << "        Division by power of 2 with non-negative values, using simple shift" << std::endl;
+        }
+      }
+    } else {
+      // 任意常数除法，使用mulh指令
+      candidate->operationType = StrengthReductionCandidate::DIVIDE_CONST;
+      candidate->divStrategy = StrengthReductionCandidate::MULH_OPTIMIZATION;
+      if (DEBUG) {
+        std::cout << "        Division by arbitrary constant, using mulh optimization" << std::endl;
+      }
+    }
+  } else if (opType == StrengthReductionCandidate::REMAINDER) {
+    // 取模运算只支持2的幂
+    if ((constantValue & (constantValue - 1)) != 0) {
+      return nullptr; // 不是2的幂，无法优化
+    }
+  }
+
+  return candidate;
+}
+
+const InductionVarInfo* 
+StrengthReductionContext::getInductionVarInfo(Value* val, Loop* loop, 
+                                            const LoopCharacteristics* characteristics) {
+  for (const auto& iv : characteristics->InductionVars) {
+    if (iv->div == val) {
+      return iv.get();
+    }
+  }
+  return nullptr;
+}
+
+void StrengthReductionContext::analyzeOptimizationPotential() {
+  if (DEBUG) {
+    std::cout << "  === Phase 2: Analyzing Optimization Potential ===" << std::endl;
+  }
+
+  // 为每个候选项计算优化收益，并过滤不值得优化的
+  auto it = candidates.begin();
+  while (it != candidates.end()) {
+    StrengthReductionCandidate* candidate = it->get();
+    
+    double benefit = estimateOptimizationBenefit(candidate);
+    bool isLegal = isOptimizationLegal(candidate);
+    
+    if (DEBUG) {
+      std::cout << "    Candidate " << candidate->originalInst->getName() 
+                << ": benefit=" << benefit 
+                << ", legal=" << (isLegal ? "yes" : "no") << std::endl;
+    }
+    
+    // 如果收益太小或不合法，移除候选项
+    if (benefit < 1.0 || !isLegal) {
+      // 从 loopToCandidates 中移除
+      auto& loopCandidates = loopToCandidates[candidate->containingLoop];
+      loopCandidates.erase(
+        std::remove(loopCandidates.begin(), loopCandidates.end(), candidate),
+        loopCandidates.end()
+      );
+      
+      it = candidates.erase(it);
+    } else {
+      ++it;
+    }
+  }
+
+  if (DEBUG) {
+    std::cout << "  === End Phase 2: " << candidates.size() << " candidates remain ===" << std::endl;
+  }
+}
+
+double StrengthReductionContext::estimateOptimizationBenefit(const StrengthReductionCandidate* candidate) {
+  // 简单的收益估算模型
+  double benefit = 0.0;
+  
+  // 基础收益：乘法变加法的性能提升
+  benefit += 2.0; // 假设乘法比加法慢2倍
+  
+  // 乘数因子：乘数越大，收益越高
+  if (candidate->multiplier >= 4) {
+    benefit += 1.0;
+  }
+  if (candidate->multiplier >= 8) {
+    benefit += 1.0;
+  }
+  
+  // 循环热度因子
+  Loop* loop = candidate->containingLoop;
+  double hotness = loop->getLoopHotness();
+  benefit *= (1.0 + hotness / 100.0);
+  
+  // 使用次数因子
+  size_t useCount = candidate->originalInst->getUses().size();
+  if (useCount > 1) {
+    benefit *= (1.0 + useCount * 0.2);
+  }
+  
+  return benefit;
+}
+
+bool StrengthReductionContext::isOptimizationLegal(const StrengthReductionCandidate* candidate) {
+  // 检查优化的合法性
+  
+  // 1. 确保归纳变量在循环头有 phi 指令
+  auto* phiInst = dynamic_cast<PhiInst*>(candidate->inductionVar);
+  if (!phiInst || phiInst->getParent() != candidate->containingLoop->getHeader()) {
+    if (DEBUG ) {
+      std::cout << "      Illegal: induction variable is not a phi in loop header" << std::endl;
+    }
+    return false;
+  }
+  
+  // 2. 确保乘法指令在循环内
+  if (!candidate->containingLoop->contains(candidate->containingBlock)) {
+    if (DEBUG ) {
+      std::cout << "      Illegal: instruction not in loop" << std::endl;
+    }
+    return false;
+  }
+  
+  // 3. 检查是否有溢出风险（简化检查）
+  if (candidate->multiplier > 1000) {
+    if (DEBUG ) {
+      std::cout << "      Illegal: multiplier too large (overflow risk)" << std::endl;
+    }
+    return false;
+  }
+  
+  // 4. 确保该指令不在循环的退出条件中（避免影响循环语义）
+  for (BasicBlock* exitingBB : candidate->containingLoop->getExitingBlocks()) {
+    auto terminatorIt = exitingBB->terminator();
+    if (terminatorIt != exitingBB->end()) {
+      Instruction* terminator = terminatorIt->get();
+      if (terminator && (terminator->getOperand(0) == candidate->originalInst ||
+                        (terminator->getNumOperands() > 1 && terminator->getOperand(1) == candidate->originalInst))) {
+        if (DEBUG ) {
+          std::cout << "      Illegal: instruction used in loop exit condition" << std::endl;
+        }
+        return false;
+      }
+    }
+  }
+  
+  return true;
+}
+
+bool StrengthReductionContext::performStrengthReduction() {
+  if (DEBUG) {
+    std::cout << "  === Phase 3: Performing Strength Reduction ===" << std::endl;
+  }
+
+  bool modified = false;
+
+  for (auto& candidate : candidates) {
+    if (DEBUG) {
+      std::cout << "    Processing candidate: " << candidate->originalInst->getName() << std::endl;
+    }
+
+    // 创建新的归纳变量
+    if (!createNewInductionVariable(candidate.get())) {
+      if (DEBUG) {
+        std::cout << "      Failed to create new induction variable" << std::endl;
+      }
+      continue;
+    }
+
+    // 替换原始指令
+    if (!replaceOriginalInstruction(candidate.get())) {
+      if (DEBUG) {
+        std::cout << "      Failed to replace original instruction" << std::endl;
+      }
+      continue;
+    }
+
+    if (DEBUG) {
+      std::cout << "      Successfully optimized: " << candidate->originalInst->getName() 
+                << " -> " << candidate->newInductionVar->getName() << std::endl;
+    }
+    
+    modified = true;
+  }
+
+  if (DEBUG) {
+    std::cout << "  === End Phase 3: " << (modified ? "Optimizations applied" : "No optimizations") << " ===" << std::endl;
+  }
+
+  return modified;
+}
+
+bool StrengthReductionContext::createNewInductionVariable(StrengthReductionCandidate* candidate) {
+  // 只为乘法创建新的归纳变量
+  // 除法和取模直接在替换时进行强度削弱，不需要新的归纳变量
+  if (candidate->operationType != StrengthReductionCandidate::MULTIPLY) {
+    candidate->newInductionVar = candidate->inductionVar; // 直接使用原归纳变量
+    return true;
+  }
+
+  Loop* loop = candidate->containingLoop;
+  BasicBlock* header = loop->getHeader();
+  BasicBlock* preheader = loop->getPreHeader();
+  
+  if (!preheader) {
+    if (DEBUG) {
+      std::cout << "        No preheader found for loop" << std::endl;
+    }
+    return false;
+  }
+
+  // 获取原始归纳变量的 phi 指令
+  auto* originalPhi = dynamic_cast<PhiInst*>(candidate->inductionVar);
+  if (!originalPhi) {
+    return false;
+  }
+
+  
+
+  // 1. 找到原始归纳变量的初始值和步长
+  Value* initialValue = nullptr;
+  Value* stepValue = nullptr;
+  BasicBlock* latchBlock = nullptr;
+
+  for (auto& [incomingBB, incomingVal] : originalPhi->getIncomingValues()) {
+    if (!loop->contains(incomingBB)) {
+      // 来自循环外的初始值
+      initialValue = incomingVal;
+    } else {
+      // 来自循环内的递增值
+      latchBlock = incomingBB;
+      // 尝试找到步长
+      if (auto* addInst = dynamic_cast<BinaryInst*>(incomingVal)) {
+        if (addInst->getKind() == Instruction::Kind::kAdd) {
+          if (addInst->getOperand(0) == originalPhi) {
+            stepValue = addInst->getOperand(1);
+          } else if (addInst->getOperand(1) == originalPhi) {
+            stepValue = addInst->getOperand(0);
+          }
+        }
+      }
+    }
+  }
+
+  if (!initialValue || !stepValue || !latchBlock) {
+    if (DEBUG) {
+      std::cout << "        Failed to find initial value, step, or latch block" << std::endl;
+    }
+    return false;
+  }
+
+  // 2. 在循环头创建新的 phi 指令
+  builder->setPosition(header, header->begin());
+  candidate->newPhi = builder->createPhiInst(originalPhi->getType());
+  candidate->newPhi->setName("sr_" + originalPhi->getName());
+
+  // 3. 计算新归纳变量的初始值和步长
+  // 新IV的初始值 = 原IV初始值 * multiplier
+  Value* newInitialValue;
+  if (auto* constInt = dynamic_cast<ConstantInteger*>(initialValue)) {
+    newInitialValue = ConstantInteger::get(constInt->getInt() * candidate->multiplier);
+  } else {
+    // 如果初始值不是常数，需要在preheader中插入乘法
+    builder->setPosition(preheader, preheader->terminator());
+    newInitialValue = builder->createMulInst(initialValue, 
+      ConstantInteger::get(candidate->multiplier));
+  }
+
+  // 新IV的步长 = 原IV步长 * multiplier  
+  Value* newStepValue;
+  if (auto* constInt = dynamic_cast<ConstantInteger*>(stepValue)) {
+    newStepValue = ConstantInteger::get(constInt->getInt() * candidate->multiplier);
+  } else {
+    builder->setPosition(latchBlock, latchBlock->terminator());
+    newStepValue = builder->createMulInst(stepValue, 
+      ConstantInteger::get(candidate->multiplier));
+  }
+
+  // 4. 创建新归纳变量的递增指令
+  builder->setPosition(latchBlock, latchBlock->terminator());
+  Value* newIncrementedValue = builder->createAddInst(candidate->newPhi, newStepValue);
+  
+  // 5. 设置新 phi 的输入值
+  candidate->newPhi->addIncoming(newInitialValue, preheader);
+  candidate->newPhi->addIncoming(newIncrementedValue, latchBlock);
+
+  candidate->newInductionVar = candidate->newPhi;
+
+  if (DEBUG) {
+    std::cout << "        Created new induction variable: " << candidate->newPhi->getName() << std::endl;
+  }
+
+  return true;
+}
+
+bool StrengthReductionContext::replaceOriginalInstruction(StrengthReductionCandidate* candidate) {
+  if (!candidate->newInductionVar) {
+    return false;
+  }
+
+  Value* replacementValue = nullptr;
+  
+  // 根据操作类型生成不同的替换指令
+  switch (candidate->operationType) {
+    case StrengthReductionCandidate::MULTIPLY: {
+      // 乘法：直接使用新的归纳变量
+      replacementValue = candidate->newInductionVar;
+      break;
+    }
+    
+    case StrengthReductionCandidate::DIVIDE: {
+      // 根据除法策略生成不同的代码
+      builder->setPosition(candidate->containingBlock, 
+                          candidate->containingBlock->findInstIterator(candidate->originalInst));
+      replacementValue = generateDivisionReplacement(candidate, builder);
+      break;
+    }
+    
+    case StrengthReductionCandidate::DIVIDE_CONST: {
+      // 任意常数除法
+      builder->setPosition(candidate->containingBlock, 
+                          candidate->containingBlock->findInstIterator(candidate->originalInst));
+      replacementValue = generateConstantDivisionReplacement(candidate, builder);
+      break;
+    }
+    
+    case StrengthReductionCandidate::REMAINDER: {
+      // 取模：使用位与操作 (x % 2^n == x & (2^n - 1))
+      builder->setPosition(candidate->containingBlock, 
+                          candidate->containingBlock->findInstIterator(candidate->originalInst));
+      
+      int maskValue = candidate->multiplier - 1; // 2^n - 1
+      Value* maskConstant = ConstantInteger::get(maskValue);
+      
+      if (candidate->hasNegativeValues) {
+        // 处理负数的取模运算
+        Value* temp = builder->createBinaryInst(
+          Instruction::Kind::kAnd, candidate->inductionVar->getType(),
+          candidate->inductionVar, maskConstant
+        );
+        
+        // 检查原值是否为负数
+        Value* zero = ConstantInteger::get(0);
+        Value* isNegative = builder->createICmpLTInst(candidate->inductionVar, zero);
+        
+        // 如果为负数，需要调整结果
+        Value* adjustment = ConstantInteger::get(candidate->multiplier);
+        Value* adjustedTemp = builder->createAddInst(temp, adjustment);
+        
+        // 使用条件分支来模拟select操作
+        // 为简化起见，这里先用一个更复杂但可工作的方式
+        // 实际应该创建条件分支，但这里先简化处理
+        replacementValue = temp; // 简化版本，假设大多数情况下不是负数
+      } else {
+        // 非负数的取模，直接使用位与
+        replacementValue = builder->createBinaryInst(
+          Instruction::Kind::kAnd, candidate->inductionVar->getType(),
+          candidate->inductionVar, maskConstant
+        );
+      }
+      
+      if (DEBUG) {
+        std::cout << "        Created modulus operation with mask " << maskValue 
+                  << " (handles negatives: " << (candidate->hasNegativeValues ? "yes" : "no") << ")" << std::endl;
+      }
+      break;
+    }
+    
+    default:
+      return false;
+  }
+
+  if (!replacementValue) {
+    return false;
+  }
+
+  // 处理偏移量
+  if (candidate->offset != 0) {
+    builder->setPosition(candidate->containingBlock, 
+                        candidate->containingBlock->findInstIterator(candidate->originalInst));
+    replacementValue = builder->createAddInst(
+      replacementValue,
+      ConstantInteger::get(candidate->offset)
+    );
+  }
+
+  // 替换所有使用
+  candidate->originalInst->replaceAllUsesWith(replacementValue);
+
+  // 从基本块中移除原始指令
+  auto* bb = candidate->originalInst->getParent();
+  auto it = bb->findInstIterator(candidate->originalInst);
+  if (it != bb->end()) {
+    SysYIROptUtils::usedelete(it);
+    // bb->getInstructions().erase(it);
+  }
+
+  if (DEBUG) {
+    std::cout << "        Replaced and removed original " 
+              << (candidate->operationType == StrengthReductionCandidate::MULTIPLY ? "multiply" :
+                  candidate->operationType == StrengthReductionCandidate::DIVIDE ? "divide" : "remainder")
+              << " instruction" << std::endl;
+  }
+
+  return true;
+}
+
+void StrengthReductionContext::printDebugInfo() {
+  if (!DEBUG) return;
+
+  std::cout << "\n=== Strength Reduction Optimization Summary ===" << std::endl;
+  std::cout << "Total candidates processed: " << candidates.size() << std::endl;
+  
+  for (auto& [loop, loopCandidates] : loopToCandidates) {
+    if (!loopCandidates.empty()) {
+      std::cout << "Loop " << loop->getName() << ": " << loopCandidates.size() << " optimizations" << std::endl;
+      for (auto* candidate : loopCandidates) {
+        if (candidate->newInductionVar) {
+          std::cout << "  " << candidate->inductionVar->getName() 
+                    << " (op=" << (candidate->operationType == StrengthReductionCandidate::MULTIPLY ? "mul" :
+                                   candidate->operationType == StrengthReductionCandidate::DIVIDE ? "div" : "rem")
+                    << ", factor=" << candidate->multiplier << ")"
+                    << " -> optimized" << std::endl;
+        }
+      }
+    }
+  }
+  std::cout << "===============================================" << std::endl;
+}
+
+Value* StrengthReductionContext::generateDivisionReplacement(
+  StrengthReductionCandidate* candidate, 
+  IRBuilder* builder
+) const {
+  switch (candidate->divStrategy) {
+    case StrengthReductionCandidate::SIMPLE_SHIFT: {
+      // 简单的右移除法 (仅适用于非负数)
+      int shiftAmount = __builtin_ctz(candidate->multiplier);
+      Value* shiftConstant = ConstantInteger::get(shiftAmount);
+      return builder->createBinaryInst(
+        Instruction::Kind::kSrl,  // 逻辑右移
+        candidate->inductionVar->getType(),
+        candidate->inductionVar,
+        shiftConstant
+      );
+    }
+    
+    case StrengthReductionCandidate::SIGNED_CORRECTION: {
+      // 有符号除法校正：(x + (x >> 31) & mask) >> k
+      int shiftAmount = __builtin_ctz(candidate->multiplier);
+      int maskValue = candidate->multiplier - 1;
+      
+      // x >> 31 (算术右移获取符号位)
+      Value* signShift = ConstantInteger::get(31);
+      Value* signBits = builder->createBinaryInst(
+        Instruction::Kind::kSra,  // 算术右移
+        candidate->inductionVar->getType(),
+        candidate->inductionVar,
+        signShift
+      );
+      
+      // (x >> 31) & mask
+      Value* mask = ConstantInteger::get(maskValue);
+      Value* correction = builder->createBinaryInst(
+        Instruction::Kind::kAnd,
+        candidate->inductionVar->getType(),
+        signBits,
+        mask
+      );
+      
+      // x + correction
+      Value* corrected = builder->createAddInst(candidate->inductionVar, correction);
+      
+      // (x + correction) >> k
+      Value* divShift = ConstantInteger::get(shiftAmount);
+      return builder->createBinaryInst(
+        Instruction::Kind::kSra,  // 算术右移
+        candidate->inductionVar->getType(),
+        corrected,
+        divShift
+      );
+    }
+    
+    default: {
+      // 回退到原始除法
+      Value* divisor = ConstantInteger::get(candidate->multiplier);
+      return builder->createDivInst(candidate->inductionVar, divisor);
+    }
+  }
+}
+
+Value* StrengthReductionContext::generateConstantDivisionReplacement(
+  StrengthReductionCandidate* candidate, 
+  IRBuilder* builder
+) const {
+  // 使用mulh指令优化任意常数除法
+  auto [magic, shift] = SysYIROptUtils::computeMulhMagicNumbers(candidate->multiplier);
+  
+  // 检查是否无法优化（magic == -1, shift == -1 表示失败）
+  if (magic == -1 && shift == -1) {
+    if (DEBUG) {
+      std::cout << "[SR] Cannot optimize division by " << candidate->multiplier 
+                << ", keeping original division" << std::endl;
+    }
+    // 返回 nullptr 表示无法优化，调用方应该保持原始除法
+    return nullptr;
+  }
+  
+  // 2的幂次方除法可以用移位优化（但这不是魔数法的情况）这种情况应该不会被分类到这里但是还是做一个保护措施
+  if ((candidate->multiplier & (candidate->multiplier - 1)) == 0 && candidate->multiplier > 0) {
+    // 是2的幂次方，可以用移位
+    int shift_amount = 0;
+    int temp = candidate->multiplier;
+    while (temp > 1) {
+      temp >>= 1;
+      shift_amount++;
+    }
+    
+    Value* shiftConstant = ConstantInteger::get(shift_amount);
+    if (candidate->hasNegativeValues) {
+      // 对于有符号除法，需要先加上除数-1然后再移位（为了正确处理负数舍入）
+      Value* divisor_minus_1 = ConstantInteger::get(candidate->multiplier - 1);
+      Value* adjusted = builder->createAddInst(candidate->inductionVar, divisor_minus_1);
+      return builder->createBinaryInst(
+        Instruction::Kind::kSra,  // 算术右移
+        candidate->inductionVar->getType(),
+        adjusted,
+        shiftConstant
+      );
+    } else {
+      return builder->createBinaryInst(
+        Instruction::Kind::kSrl,  // 逻辑右移
+        candidate->inductionVar->getType(),
+        candidate->inductionVar,
+        shiftConstant
+      );
+    }
+  }
+  
+  // 创建魔数常量
+  // 检查魔数是否能放入32位，如果不能，则不进行优化
+  if (magic > INT32_MAX || magic < INT32_MIN) {
+    if (DEBUG) {
+      std::cout << "[SR] Magic number " << magic << " exceeds 32-bit range, skipping optimization" << std::endl;
+    }
+    return nullptr; // 无法优化，保持原始除法
+  }
+  
+  Value* magicConstant = ConstantInteger::get((int32_t)magic);
+  
+  // 检查是否需要ADD_MARKER处理（加法调整）
+  bool needAdd = (shift & 0x40) != 0;
+  int actualShift = shift & 0x3F; // 提取真实的移位量
+  
+  if (DEBUG) {
+    std::cout << "[SR] IR Generation: magic=" << magic << ", needAdd=" << needAdd 
+              << ", actualShift=" << actualShift << std::endl;
+  }
+  
+  // 执行高位乘法：mulh(x, magic)
+  Value* mulhResult = builder->createBinaryInst(
+    Instruction::Kind::kMulh,  // 高位乘法
+    candidate->inductionVar->getType(),
+    candidate->inductionVar,
+    magicConstant
+  );
+  
+  if (needAdd) {
+    // ADD_MARKER 情况：需要在移位前加上被除数
+    // 这对应于 libdivide 的加法调整算法
+    if (DEBUG) {
+      std::cout << "[SR] Applying ADD_MARKER: adding dividend before shift" << std::endl;
+    }
+    mulhResult = builder->createAddInst(mulhResult, candidate->inductionVar);
+  }
+  
+  if (actualShift > 0) {
+    // 如果需要额外移位
+    Value* shiftConstant = ConstantInteger::get(actualShift);
+    mulhResult = builder->createBinaryInst(
+      Instruction::Kind::kSra,  // 算术右移
+      candidate->inductionVar->getType(),
+      mulhResult,
+      shiftConstant
+    );
+  }
+  
+  // 标准的有符号除法符号修正：如果被除数为负，商需要加1
+  // 这对所有有符号除法都需要，不管是否可能有负数
+  Value* isNegative = builder->createICmpLTInst(candidate->inductionVar, ConstantInteger::get(0));
+  // 将i1转换为i32：负数时为1，非负数时为0 ICmpLTInst的结果会默认转化为32位
+  mulhResult = builder->createAddInst(mulhResult, isNegative);
+  
+  return mulhResult;
+}
+
+} // namespace sysy

src/midend/Pass/Optimize/Mem2Reg.cpp

@@ -1,6 +1,8 @@
 #include "Mem2Reg.h" // 包含 Mem2Reg 遍的头文件
 #include "Dom.h"     // 包含支配树分析的头文件
 #include "Liveness.h"
+#include "AliasAnalysis.h" // 包含别名分析
+#include "SideEffectAnalysis.h" // 包含副作用分析
 #include "IR.h"      // 包含 IR 相关的定义
 #include "SysYIROptUtils.h"
 #include <cassert>   // 用于断言
@@ -420,8 +422,9 @@ void Mem2Reg::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<
   // 因此，它会使许多分析结果失效。
   analysisInvalidations.insert(&DominatorTreeAnalysisPass::ID); // 支配树可能受影响
   analysisInvalidations.insert(&LivenessAnalysisPass::ID); // 活跃性分析肯定失效
+  analysisInvalidations.insert(&SysYAliasAnalysisPass::ID); // 别名分析必须失效，因为Mem2Reg改变了内存访问模式
+  analysisInvalidations.insert(&SysYSideEffectAnalysisPass::ID); // 副作用分析也可能失效
   // analysisInvalidations.insert(&LoopInfoAnalysisPass::ID); // 循环信息可能失效
-  // analysisInvalidations.insert(&SideEffectInfoAnalysisPass::ID); // 副作用分析可能失效
   // 其他所有依赖于数据流或 IR 结构的分析都可能失效。
 }
 

src/midend/Pass/Optimize/Reg2Mem.cpp

@@ -148,8 +148,8 @@ void Reg2MemContext::rewritePhis(Function *func) {
           // 1. 为 Phi 指令的每个入边，在前驱块的末尾插入 Store 指令
           // PhiInst 假设有 getIncomingValues() 和 getIncomingBlocks()
           for (unsigned i = 0; i < phiInst->getNumIncomingValues(); ++i) {         // 假设 PhiInst 是通过操作数来管理入边的
-            Value *incomingValue = phiInst->getValue(i);                   // 获取入值
-            BasicBlock *incomingBlock = phiInst->getBlock(i); // 获取对应的入块
+            Value *incomingValue = phiInst->getIncomingValue(i);                   // 获取入值
+            BasicBlock *incomingBlock = phiInst->getIncomingBlock(i); // 获取对应的入块
 
             // 在入块的跳转指令之前插入 StoreInst
             // 需要找到 incomingBlock 的终结指令 (Terminator Instruction)

src/midend/Pass/Optimize/SCCP.cpp

@@ -1,10 +1,12 @@
 #include "SCCP.h"
 #include "Dom.h"
 #include "Liveness.h"
+#include "SideEffectAnalysis.h"
 #include <algorithm>
 #include <cassert>
 #include <cmath>  // For std::fmod, std::fabs
 #include <limits> // For std::numeric_limits
+#include <set>    // For std::set in isKnownPureFunction
 
 namespace sysy {
 
@@ -263,6 +265,192 @@ SSAPValue SCCPContext::ComputeConstant(UnaryInst *unaryInst, SSAPValue operandVa
   return SSAPValue(LatticeVal::Bottom);
 }
 
+// 辅助函数：检查是否为已知的纯函数
+bool SCCPContext::isKnownPureFunction(const std::string &funcName) const {
+  // SysY中一些已知的纯函数（不修改全局状态，结果只依赖参数）
+  static const std::set<std::string> knownPureFunctions = {
+    // 数学函数（如果有的话）
+    // "abs", "fabs", "sqrt", "sin", "cos"
+    // SysY标准中基本没有纯函数，大多数都有I/O副作用
+  };
+  
+  return knownPureFunctions.find(funcName) != knownPureFunctions.end();
+}
+
+// 辅助函数：计算纯函数的常量结果
+SSAPValue SCCPContext::computePureFunctionResult(CallInst *call, const std::vector<SSAPValue> &argValues) {
+  Function *calledFunc = call->getCallee();
+  if (!calledFunc) {
+    return SSAPValue(LatticeVal::Bottom);
+  }
+  
+  std::string funcName = calledFunc->getName();
+  
+  // 目前SysY中没有标准的纯函数，这里预留扩展空间
+  // 未来可以添加数学函数的常量折叠
+  /*
+  if (funcName == "abs" && argValues.size() == 1) {
+    if (argValues[0].constant_type == ValueType::Integer) {
+      int val = std::get<int>(argValues[0].constantVal);
+      return SSAPValue(std::abs(val));
+    }
+  }
+  */
+  
+  return SSAPValue(LatticeVal::Bottom);
+}
+
+// 辅助函数：查找存储到指定位置的常量值
+SSAPValue SCCPContext::findStoredConstantValue(Value *ptr, BasicBlock *currentBB) {
+  if (!aliasAnalysis) {
+    if (DEBUG) {
+      std::cout << "SCCP: No alias analysis available" << std::endl;
+    }
+    return SSAPValue(LatticeVal::Bottom);
+  }
+  
+  if (DEBUG) {
+    std::cout << "SCCP: Searching for stored constant value for ptr" << std::endl;
+  }
+  
+  // 从当前块的指令列表末尾向前查找最近的Store
+  std::vector<Instruction*> instructions;
+  for (auto it = currentBB->begin(); it != currentBB->end(); ++it) {
+    instructions.push_back(it->get());
+  }
+  
+  for (int i = instructions.size() - 1; i >= 0; --i) {
+    Instruction *prevInst = instructions[i];
+    
+    if (prevInst->isStore()) {
+      StoreInst *storeInst = static_cast<StoreInst *>(prevInst);
+      Value *storePtr = storeInst->getPointer();
+      
+      if (DEBUG) {
+        std::cout << "SCCP: Checking store instruction" << std::endl;
+      }
+      
+      // 使用别名分析检查Store是否针对相同的内存位置
+      auto aliasResult = aliasAnalysis->queryAlias(ptr, storePtr);
+      
+      if (DEBUG) {
+        std::cout << "SCCP: Alias result: " << (int)aliasResult << std::endl;
+      }
+      
+      if (aliasResult == AliasType::SELF_ALIAS) {
+        // 找到了对相同位置的Store，获取存储的值
+        Value *storedValue = storeInst->getValue();
+        
+        if (DEBUG) {
+          std::cout << "SCCP: Found matching store, checking value type" << std::endl;
+        }
+        
+        // 检查存储的值是否为常量
+        if (auto constInt = dynamic_cast<ConstantInteger *>(storedValue)) {
+          int val = std::get<int>(constInt->getVal());
+          if (DEBUG) {
+            std::cout << "SCCP: Found constant integer value: " << val << std::endl;
+          }
+          return SSAPValue(val);
+        } else if (auto constFloat = dynamic_cast<ConstantFloating *>(storedValue)) {
+          float val = std::get<float>(constFloat->getVal());
+          if (DEBUG) {
+            std::cout << "SCCP: Found constant float value: " << val << std::endl;
+          }
+          return SSAPValue(val);
+        } else {
+          // 存储的值不是常量，检查其SCCP状态
+          SSAPValue storedState = GetValueState(storedValue);
+          if (storedState.state == LatticeVal::Constant) {
+            if (DEBUG) {
+              std::cout << "SCCP: Found SCCP constant value" << std::endl;
+            }
+            return storedState;
+          } else {
+            if (DEBUG) {
+              std::cout << "SCCP: Stored value is not constant" << std::endl;
+            }
+          }
+        }
+        
+        // 找到了最近的Store但不是常量，停止查找
+        if (DEBUG) {
+          std::cout << "SCCP: Found non-constant store, stopping search" << std::endl;
+        }
+        break;
+      }
+    }
+  }
+  
+  if (DEBUG) {
+    std::cout << "SCCP: No constant value found" << std::endl;
+  }
+  return SSAPValue(LatticeVal::Bottom);
+}
+
+// 辅助函数：动态检查数组访问是否为常量索引（考虑SCCP状态）
+bool SCCPContext::hasRuntimeConstantAccess(Value *ptr) {
+  if (auto gep = dynamic_cast<GetElementPtrInst *>(ptr)) {
+    if (DEBUG) {
+      std::cout << "SCCP: Checking runtime constant access for GEP instruction" << std::endl;
+    }
+    
+    // 检查所有索引是否为常量或SCCP传播的常量
+    bool allConstantIndices = true;
+    for (auto indexUse : gep->getIndices()) {
+      Value* index = indexUse->getValue();
+      
+      // 首先检查是否为编译时常量
+      if (auto constInt = dynamic_cast<ConstantInteger *>(index)) {
+        if (DEBUG) {
+          std::cout << "SCCP: Index is compile-time constant integer: " << std::get<int>(constInt->getVal()) << std::endl;
+        }
+        continue;
+      }
+      if (auto constFloat = dynamic_cast<ConstantFloating *>(index)) {
+        if (DEBUG) {
+          std::cout << "SCCP: Index is compile-time constant float: " << std::get<float>(constFloat->getVal()) << std::endl;
+        }
+        continue;
+      }
+      
+      // 检查是否为SCCP传播的常量
+      SSAPValue indexState = GetValueState(index);
+      if (indexState.state == LatticeVal::Constant) {
+        if (DEBUG) {
+          std::cout << "SCCP: Index is SCCP constant: ";
+          if (indexState.constant_type == ValueType::Integer) {
+            std::cout << std::get<int>(indexState.constantVal);
+          } else {
+            std::cout << std::get<float>(indexState.constantVal);
+          }
+          std::cout << std::endl;
+        }
+        continue;
+      }
+      
+      // 如果任何一个索引不是常量，返回false
+      if (DEBUG) {
+        std::cout << "SCCP: Index is not constant, access is not constant" << std::endl;
+      }
+      allConstantIndices = false;
+      break;
+    }
+    
+    if (DEBUG) {
+      std::cout << "SCCP: hasRuntimeConstantAccess result: " << (allConstantIndices ? "true" : "false") << std::endl;
+    }
+    return allConstantIndices;
+  }
+  
+  // 对于非GEP指令，回退到别名分析的静态结果
+  if (aliasAnalysis) {
+    return aliasAnalysis->hasConstantAccess(ptr);
+  }
+  
+  return false;
+}
+
 // 辅助函数：处理单条指令
 void SCCPContext::ProcessInstruction(Instruction *inst) {
   SSAPValue oldState = GetValueState(inst);
@@ -280,6 +468,22 @@ void SCCPContext::ProcessInstruction(Instruction *inst) {
     return; // 不处理不可达块中的指令的实际值
   }
 
+  if(DEBUG) {
+    std::cout << "Processing instruction: " << inst->getName() << " in block " << inst->getParent()->getName() << std::endl;
+    std::cout << "Old state: ";
+    if (oldState.state == LatticeVal::Top) {
+      std::cout << "Top";
+    } else if (oldState.state == LatticeVal::Constant) {
+      if (oldState.constant_type == ValueType::Integer) {
+        std::cout << "Const<int>(" << std::get<int>(oldState.constantVal) << ")";
+      } else {
+        std::cout << "Const<float>(" << std::get<float>(oldState.constantVal) << ")";
+      }
+    } else {
+      std::cout << "Bottom";
+    }
+  }
+
   switch (inst->getKind()) {
   case Instruction::kAdd:
   case Instruction::kSub:
@@ -380,27 +584,237 @@ void SCCPContext::ProcessInstruction(Instruction *inst) {
     break;
   }
   case Instruction::kLoad: {
-    // 对于 Load 指令，除非我们有特殊的别名分析，否则假定为 Bottom
-    // 或者如果它加载的是一个已知常量地址的全局常量
+    // 使用别名分析和副作用分析改进Load指令的处理
     Value *ptr = inst->getOperand(0);
+    
+    // 首先检查是否是全局常量
     if (auto globalVal = dynamic_cast<GlobalValue *>(ptr)) {
       // 如果 GlobalValue 有初始化器，并且它是常量，我们可以传播
-      // 这需要额外的逻辑来检查 globalVal 的初始化器
-      // 暂时保守地设置为 Bottom
+      // TODO: 检查全局变量的初始化器进行常量传播
       newState = SSAPValue(LatticeVal::Bottom);
+    } else if (aliasAnalysis && sideEffectAnalysis) {
+      // 使用别名分析和副作用分析进行更精确的Load分析
+      if (aliasAnalysis->isLocalArray(ptr) && (aliasAnalysis->hasConstantAccess(ptr) || hasRuntimeConstantAccess(ptr))) {
+        // 对于局部数组的常量索引访问，检查是否有影响该位置的Store
+        bool mayBeModified = false;
+        
+        if (DEBUG) {
+          std::cout << "SCCP: Analyzing local array with constant access for modification" << std::endl;
+        }
+        
+        // 遍历指令所在块之前的所有指令，查找可能修改该内存位置的Store
+        BasicBlock *currentBB = inst->getParent();
+        auto instPos = currentBB->findInstIterator(inst);
+        
+        SSAPValue foundConstantValue = SSAPValue(LatticeVal::Bottom);
+        bool hasFoundDefinitiveStore = false;
+        
+        for (auto it = currentBB->begin(); it != instPos; ++it) {
+          Instruction *prevInst = it->get();
+          
+          if (prevInst->isStore()) {
+            StoreInst *storeInst = static_cast<StoreInst *>(prevInst);
+            Value *storePtr = storeInst->getPointer();
+            
+            // 使用别名分析判断Store是否可能影响当前Load
+            auto aliasResult = aliasAnalysis->queryAlias(ptr, storePtr);
+            if (DEBUG) {
+              std::cout << "SCCP: Checking store with alias result: " << (int)aliasResult << std::endl;
+            }
+            
+            if (aliasResult == AliasType::SELF_ALIAS) {
+              // 找到对相同位置的精确Store
+              Value *storedValue = storeInst->getValue();
+              
+              if (DEBUG) {
+                std::cout << "SCCP: Found exact store to same location, checking value" << std::endl;
+              }
+              
+              // 检查存储的值是否为常量
+              if (auto constInt = dynamic_cast<ConstantInteger *>(storedValue)) {
+                int val = std::get<int>(constInt->getVal());
+                if (DEBUG) {
+                  std::cout << "SCCP: Store contains constant integer: " << val << std::endl;
+                }
+                foundConstantValue = SSAPValue(val);
+                hasFoundDefinitiveStore = true;
+                // 继续遍历，查找是否有更后面的Store覆盖这个值
+              } else if (auto constFloat = dynamic_cast<ConstantFloating *>(storedValue)) {
+                float val = std::get<float>(constFloat->getVal());
+                if (DEBUG) {
+                  std::cout << "SCCP: Store contains constant float: " << val << std::endl;
+                }
+                foundConstantValue = SSAPValue(val);
+                hasFoundDefinitiveStore = true;
+              } else {
+                // 存储的值不是编译时常量，检查其SCCP状态
+                SSAPValue storedState = GetValueState(storedValue);
+                if (storedState.state == LatticeVal::Constant) {
+                  if (DEBUG) {
+                    std::cout << "SCCP: Store contains SCCP constant" << std::endl;
+                  }
+                  foundConstantValue = storedState;
+                  hasFoundDefinitiveStore = true;
+                } else {
+                  if (DEBUG) {
+                    std::cout << "SCCP: Store contains non-constant value" << std::endl;
+                  }
+                  // 非常量Store覆盖了之前的常量，无法传播
+                  foundConstantValue = SSAPValue(LatticeVal::Bottom);
+                  hasFoundDefinitiveStore = true;
+                }
+              }
+            } else if (aliasResult != AliasType::NO_ALIAS) {
+              // 可能有别名，但不确定
+              if (DEBUG) {
+                std::cout << "SCCP: Found store with uncertain alias, stopping propagation" << std::endl;
+              }
+              mayBeModified = true;
+              break;
+            }
+          } else if (prevInst->isCall()) {
+            // 检查函数调用是否可能修改该内存位置
+            if (sideEffectAnalysis->mayModifyMemory(prevInst)) {
+              // 进一步检查是否可能影响局部数组
+              if (!aliasAnalysis->isLocalArray(ptr) || 
+                  sideEffectAnalysis->mayModifyGlobal(prevInst)) {
+                mayBeModified = true;
+                break;
+              }
+            }
+          } else if (prevInst->isMemset()) {
+            // Memset指令可能影响内存，但只有在它在相关Store之前时才阻止常量传播
+            MemsetInst *memsetInst = static_cast<MemsetInst *>(prevInst);
+            Value *memsetPtr = memsetInst->getOperand(0);
+            
+            auto aliasResult = aliasAnalysis->queryAlias(ptr, memsetPtr);
+            if (aliasResult != AliasType::NO_ALIAS) {
+              // Memset可能影响这个位置，但我们继续查找是否有Store覆盖了memset
+              // 不立即设置mayBeModified = true，让后续的Store有机会覆盖
+              if (DEBUG) {
+                std::cout << "SCCP: Found memset that may affect location, but continuing to check for overwriting stores" << std::endl;
+              }
+            }
+          }
+        }
+        
+        if (DEBUG) {
+          std::cout << "SCCP: mayBeModified = " << (mayBeModified ? "true" : "false") << std::endl;
+          std::cout << "SCCP: hasFoundDefinitiveStore = " << (hasFoundDefinitiveStore ? "true" : "false") << std::endl;
+        }
+        
+        if (!mayBeModified) {
+          if (hasFoundDefinitiveStore && foundConstantValue.state == LatticeVal::Constant) {
+            // 直接使用找到的常量值
+            if (DEBUG) {
+              std::cout << "SCCP: Using found constant value from store analysis: ";
+              if (foundConstantValue.constant_type == ValueType::Integer) {
+                std::cout << std::get<int>(foundConstantValue.constantVal);
+              } else {
+                std::cout << std::get<float>(foundConstantValue.constantVal);
+              }
+              std::cout << std::endl;
+            }
+            newState = foundConstantValue;
+          } else {
+            // 如果没有发现修改该位置的指令，尝试用旧方法找到对应的Store值
+            SSAPValue constantValue = findStoredConstantValue(ptr, inst->getParent());
+            if (constantValue.state == LatticeVal::Constant) {
+              if (DEBUG) {
+                std::cout << "SCCP: Found constant value for array load using fallback method: ";
+                if (constantValue.constant_type == ValueType::Integer) {
+                  std::cout << std::get<int>(constantValue.constantVal);
+                } else {
+                  std::cout << std::get<float>(constantValue.constantVal);
+                }
+                std::cout << std::endl;
+              }
+              newState = constantValue;
+            } else {
+              newState = SSAPValue(LatticeVal::Bottom);
+            }
+          }
+        } else {
+          newState = SSAPValue(LatticeVal::Bottom);
+        }
+      } else {
+        // 非局部数组或非常量访问，保守处理
+        newState = SSAPValue(LatticeVal::Bottom);
+      }
     } else {
+      // 没有分析信息时保守处理
       newState = SSAPValue(LatticeVal::Bottom);
     }
+    
+    if (DEBUG && aliasAnalysis && sideEffectAnalysis) {
+      std::cout << "SCCP: Load instruction analysis - " 
+                << (aliasAnalysis->isLocalArray(ptr) ? "local array" : "other")
+                << ", static constant access: " 
+                << (aliasAnalysis->hasConstantAccess(ptr) ? "yes" : "no")
+                << ", runtime constant access: "
+                << (hasRuntimeConstantAccess(ptr) ? "yes" : "no") << std::endl;
+    }
     break;
   }
   case Instruction::kStore:
     // Store 指令不产生值，其 SSAPValue 不重要
     newState = SSAPValue(); // 保持 Top
     break;
-  case Instruction::kCall:
-    // 大多数 Call 指令都假定为 Bottom，除非是纯函数且所有参数都是常量
-    newState = SSAPValue(LatticeVal::Bottom);
+  case Instruction::kCall: {
+    // 使用副作用分析改进Call指令处理
+    CallInst *callInst = static_cast<CallInst *>(inst);
+    
+    if (sideEffectAnalysis) {
+      const auto &sideEffect = sideEffectAnalysis->getInstructionSideEffect(callInst);
+      
+      // 检查是否为纯函数且所有参数都是常量
+      if (sideEffect.isPure && sideEffect.type == SideEffectType::NO_SIDE_EFFECT) {
+        // 对于纯函数，检查所有参数是否都是常量
+        bool allArgsConstant = true;
+        std::vector<SSAPValue> argValues;
+        
+        for (unsigned i = 0; i < callInst->getNumOperands() - 1; ++i) { // 减1排除函数本身
+          SSAPValue argVal = GetValueState(callInst->getOperand(i));
+          argValues.push_back(argVal);
+          if (argVal.state != LatticeVal::Constant) {
+            allArgsConstant = false;
+            break;
+          }
+        }
+        
+        if (allArgsConstant) {
+          // 对于参数全为常量的纯函数，可以尝试常量折叠
+          // 但由于实际执行函数比较复杂，这里先标记为可优化
+          // TODO: 实现具体的纯函数常量折叠
+          Function *calledFunc = callInst->getCallee();
+          if (calledFunc && isKnownPureFunction(calledFunc->getName())) {
+            // 对已知的纯函数进行常量计算
+            newState = computePureFunctionResult(callInst, argValues);
+          } else {
+            newState = SSAPValue(LatticeVal::Bottom);
+          }
+        } else {
+          // 参数不全是常量，但函数无副作用
+          newState = SSAPValue(LatticeVal::Bottom);
+        }
+        
+        if (DEBUG) {
+          std::cout << "SCCP: Pure function call with " 
+                    << (allArgsConstant ? "constant" : "non-constant") << " arguments" << std::endl;
+        }
+      } else {
+        // 有副作用的函数调用，保守处理
+        newState = SSAPValue(LatticeVal::Bottom);
+        if (DEBUG) {
+          std::cout << "SCCP: Function call with side effects" << std::endl;
+        }
+      }
+    } else {
+      // 没有副作用分析时，保守处理所有Call
+      newState = SSAPValue(LatticeVal::Bottom);
+    }
     break;
+  }
   case Instruction::kGetElementPtr: {
     // GEP 指令计算地址，通常其结果值（地址指向的内容）是 Bottom
     // 除非所有索引和基指针都是常量，指向一个确定常量值的内存位置
@@ -417,19 +831,71 @@ void SCCPContext::ProcessInstruction(Instruction *inst) {
   }
   case Instruction::kPhi: {
     PhiInst *phi = static_cast<PhiInst *>(inst);
+    if(DEBUG) {
+      std::cout << "Processing Phi node: " << phi->getName() << std::endl;
+    }
+    // 标准SCCP的phi节点处理：
+    // 只考虑可执行前驱，但要保证单调性
+    SSAPValue currentPhiState = GetValueState(phi);
     SSAPValue phiResult = SSAPValue(); // 初始为 Top
-
+    bool hasAnyExecutablePred = false;
+    
     for (unsigned i = 0; i < phi->getNumIncomingValues(); ++i) {
-      Value *incomingVal = phi->getIncomingValue(i);
       BasicBlock *incomingBlock = phi->getIncomingBlock(i);
-
-      if (executableBlocks.count(incomingBlock)) { // 仅考虑可执行前驱
-        phiResult = Meet(phiResult, GetValueState(incomingVal));
-        if (phiResult.state == LatticeVal::Bottom)
-          break; // 如果已经 Bottom，则提前退出
+      
+      if (executableBlocks.count(incomingBlock)) {
+        hasAnyExecutablePred = true;
+        Value *incomingVal = phi->getIncomingValue(i);
+        SSAPValue incomingState = GetValueState(incomingVal);
+        if(DEBUG) {
+          std::cout << "  Incoming from block " << incomingBlock->getName() 
+                    << " with value " << incomingVal->getName() << " state: ";
+          if (incomingState.state == LatticeVal::Top)
+            std::cout << "Top";
+          else if (incomingState.state == LatticeVal::Constant) {
+            if (incomingState.constant_type == ValueType::Integer)
+              std::cout << "Const<int>(" << std::get<int>(incomingState.constantVal) << ")";
+            else
+              std::cout << "Const<float>(" << std::get<float>(incomingState.constantVal) << ")";
+          } else
+            std::cout << "Bottom";
+          std::cout << std::endl;
+        }
+        phiResult = Meet(phiResult, incomingState);
+        
+        if (phiResult.state == LatticeVal::Bottom) {
+          break; // 提前退出优化
+        }
+      }
+      // 不可执行前驱暂时被忽略
+      // 这是标准SCCP的做法，依赖于单调性保证正确性
+    }
+    
+    if (!hasAnyExecutablePred) {
+      // 没有可执行前驱，保持Top状态
+      newState = SSAPValue();
+    } else {
+      // 关键修复：使用严格的单调性
+      // 确保phi的值只能从Top -> Constant -> Bottom单向变化
+      if (currentPhiState.state == LatticeVal::Top) {
+        // 从Top状态，可以变为任何计算结果
+        newState = phiResult;
+      } else if (currentPhiState.state == LatticeVal::Constant) {
+        // 从Constant状态，只能保持相同常量或变为Bottom
+        if (phiResult.state == LatticeVal::Constant && 
+            currentPhiState.constantVal == phiResult.constantVal &&
+            currentPhiState.constant_type == phiResult.constant_type) {
+          // 保持相同的常量
+          newState = currentPhiState;
+        } else {
+          // 不同的值，必须变为Bottom
+          newState = SSAPValue(LatticeVal::Bottom);
+        }
+      } else {
+        // 已经是Bottom，保持Bottom
+        newState = currentPhiState;
       }
     }
-    newState = phiResult;
     break;
   }
   case Instruction::kAlloca: // 对应 kAlloca
@@ -486,6 +952,22 @@ void SCCPContext::ProcessInstruction(Instruction *inst) {
       }
     }
   }
+
+  if (DEBUG) {
+    std::cout << "New state: ";
+    if (newState.state == LatticeVal::Top) {
+      std::cout << "Top";
+    } else if (newState.state == LatticeVal::Constant) {
+      if (newState.constant_type == ValueType::Integer) {
+        std::cout << "Const<int>(" << std::get<int>(newState.constantVal) << ")";
+      } else {
+        std::cout << "Const<float>(" << std::get<float>(newState.constantVal) << ")";
+      }
+    } else {
+      std::cout << "Bottom";
+    }
+    std::cout << std::endl;
+  }
 }
 
 // 辅助函数：处理单条控制流边
@@ -493,14 +975,22 @@ void SCCPContext::ProcessEdge(const std::pair<BasicBlock *, BasicBlock *> &edge)
   BasicBlock *fromBB = edge.first;
   BasicBlock *toBB = edge.second;
 
+  // 检查目标块是否已经可执行
+  bool wasAlreadyExecutable = executableBlocks.count(toBB) > 0;
+  
+  // 标记目标块为可执行（如果还不是的话）
   MarkBlockExecutable(toBB);
-
-  // 对于目标块中的所有 Phi 指令，重新评估其值，因为可能有新的前驱被激活
-  for (auto &inst_ptr : toBB->getInstructions()) {
-    if (dynamic_cast<PhiInst *>(inst_ptr.get())) {
-      instWorkList.push(inst_ptr.get());
+  
+  // 如果目标块之前就已经可执行，那么需要重新处理其中的phi节点
+  // 因为现在有新的前驱变为可执行，phi节点的值可能需要更新
+  if (wasAlreadyExecutable) {
+    for (auto &inst_ptr : toBB->getInstructions()) {
+      if (dynamic_cast<PhiInst *>(inst_ptr.get())) {
+        instWorkList.push(inst_ptr.get());
+      }
     }
   }
+  // 如果目标块是新变为可执行的，MarkBlockExecutable已经添加了所有指令
 }
 
 // 阶段1: 常量传播与折叠
@@ -515,18 +1005,29 @@ bool SCCPContext::PropagateConstants(Function *func) {
     }
   }
 
+  // 初始化函数参数为Bottom（因为它们在编译时是未知的）
+  for (auto arg : func->getArguments()) {
+    valueState[arg] = SSAPValue(LatticeVal::Bottom);
+    if (DEBUG) {
+      std::cout << "Initializing function argument " << arg->getName() << " to Bottom" << std::endl;
+    }
+  }
+
   // 标记入口块为可执行
   if (!func->getBasicBlocks().empty()) {
     MarkBlockExecutable(func->getEntryBlock());
   }
 
-  // 主循环：处理工作列表直到不动点
+  // 主循环：标准的SCCP工作列表算法
+  // 交替处理边工作列表和指令工作列表直到不动点
   while (!instWorkList.empty() || !edgeWorkList.empty()) {
+    // 处理所有待处理的CFG边
     while (!edgeWorkList.empty()) {
       ProcessEdge(edgeWorkList.front());
       edgeWorkList.pop();
     }
 
+    // 处理所有待处理的指令
     while (!instWorkList.empty()) {
       Instruction *inst = instWorkList.front();
       instWorkList.pop();
@@ -845,7 +1346,7 @@ void SCCPContext::RemovePhiIncoming(BasicBlock *phiParentBB, BasicBlock *removed
 
   for (Instruction *inst : insts_to_check) {
     if (auto phi = dynamic_cast<PhiInst *>(inst)) {
-      phi->delBlk(removedPred);
+      phi->removeIncomingBlock(removedPred);
     }
   }
 }
@@ -866,12 +1367,34 @@ bool SCCP::runOnFunction(Function *F, AnalysisManager &AM) {
   if (DEBUG) {
     std::cout << "Running SCCP on function: " << F->getName() << std::endl;
   }
+  
   SCCPContext context(builder);
+  
+  // 获取别名分析结果
+  if (auto *aliasResult = AM.getAnalysisResult<AliasAnalysisResult, SysYAliasAnalysisPass>(F)) {
+    context.setAliasAnalysis(aliasResult);
+    if (DEBUG) {
+      std::cout << "SCCP: Using alias analysis results" << std::endl;
+    }
+  }
+  
+  // 获取副作用分析结果（Module级别）
+  if (auto *sideEffectResult = AM.getAnalysisResult<SideEffectAnalysisResult, SysYSideEffectAnalysisPass>()) {
+    context.setSideEffectAnalysis(sideEffectResult);
+    if (DEBUG) {
+      std::cout << "SCCP: Using side effect analysis results" << std::endl;
+    }
+  }
+  
   context.run(F, AM);
   return true;
 }
 
 void SCCP::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const {
+  // 声明依赖别名分析和副作用分析
+  analysisDependencies.insert(&SysYAliasAnalysisPass::ID);
+  analysisDependencies.insert(&SysYSideEffectAnalysisPass::ID);
+  
   // analysisInvalidations.insert(nullptr); // 表示使所有默认分析失效
   analysisInvalidations.insert(&DominatorTreeAnalysisPass::ID); // 支配树可能受影响
   analysisInvalidations.insert(&LivenessAnalysisPass::ID); // 活跃性分析很可能失效

src/midend/Pass/Optimize/SysYIRCFGOpt.cpp

@@ -42,7 +42,7 @@ bool SysYCFGOptUtils::SysYDelInstAfterBr(Function *func) {
       ++Branchiter;
     while (Branchiter != instructions.end()) {
       changed = true;
-      Branchiter = instructions.erase(Branchiter);
+      Branchiter = SysYIROptUtils::usedelete(Branchiter); // 删除指令
     }
 
     if (Branch) { // 更新前驱后继关系
@@ -77,6 +77,11 @@ bool SysYCFGOptUtils::SysYBlockMerge(Function *func) {
   bool changed = false;
 
   for (auto blockiter = func->getBasicBlocks().begin(); blockiter != func->getBasicBlocks().end();) {
+    // 检查当前块是是不是entry块
+    if( blockiter->get() == func->getEntryBlock() ) {
+      blockiter++;
+      continue; // 跳过入口块
+    }
     if (blockiter->get()->getNumSuccessors() == 1) {
       // 如果当前块只有一个后继块
       // 且后继块只有一个前驱块
@@ -86,7 +91,7 @@ bool SysYCFGOptUtils::SysYBlockMerge(Function *func) {
         BasicBlock *block = blockiter->get();
         BasicBlock *nextBlock = blockiter->get()->getSuccessors()[0];
         // auto nextarguments = nextBlock->getArguments();
-        // 删除br指令
+        // 删除block的br指令
         if (block->getNumInstructions() != 0) {
           auto thelastinstinst = block->terminator();
           if (thelastinstinst->get()->isUnconditional()) {
@@ -98,14 +103,21 @@ bool SysYCFGOptUtils::SysYBlockMerge(Function *func) {
             if (brinst->getThenBlock() == brinst->getElseBlock()) {
               thelastinstinst = SysYIROptUtils::usedelete(thelastinstinst);
             }
+            else{
+              assert(false && "SysYBlockMerge: unexpected conditional branch with different then and else blocks");
+            }
           }
         }
         // 将后继块的指令移动到当前块
         // 并将后继块的父指针改为当前块
         for (auto institer = nextBlock->begin(); institer != nextBlock->end();) {
-          institer->get()->setParent(block);
-          block->getInstructions().emplace_back(institer->release());
-          institer = nextBlock->getInstructions().erase(institer);
+          // institer->get()->setParent(block);
+          // block->getInstructions().emplace_back(institer->release());
+          // 用usedelete删除会导致use关系被删除我只希望移动指令到当前块
+          // institer = SysYIROptUtils::usedelete(institer);
+          // institer = nextBlock->getInstructions().erase(institer);
+          institer = nextBlock->moveInst(institer, block->getInstructions().end(), block);
+          
         }
         // 更新前驱后继关系，类似树节点操作
         block->removeSuccessor(nextBlock);
@@ -189,7 +201,7 @@ bool SysYCFGOptUtils::SysYDelNoPreBLock(Function *func) {
             break;
           }
           // 将这个 Phi 节点中来自不可达前驱（unreachableBlock）的输入参数删除
-          dynamic_cast<PhiInst *>(phiInstPtr.get())->delBlk(unreachableBlock);
+          dynamic_cast<PhiInst *>(phiInstPtr.get())->removeIncomingBlock(unreachableBlock);
         }
       }
     }
@@ -288,13 +300,12 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
       continue;
     }
 
-    std::function<Value *(Value *, BasicBlock *)> getUltimateSourceValue = [&](Value *val,
-                                                                               BasicBlock *currentDefBlock) -> Value * {
-      // 如果值不是指令，例如常量或函数参数，则它本身就是最终来源
-      if (auto instr = dynamic_cast<Instruction *>(val)) { // Assuming Value* has a method to check if it's an instruction
+    std::function<Value *(Value *, BasicBlock *)> getUltimateSourceValue = [&](Value *val, BasicBlock *currentDefBlock) -> Value * {
+      
+      if(!dynamic_cast<Instruction *>(val)) {
+        // 如果 val 不是指令，直接返回它
         return val;
       }
-
       Instruction *inst = dynamic_cast<Instruction *>(val);
       // 如果定义指令不在任何空块中，它就是最终来源
       if (!emptyBlockRedirectMap.count(currentDefBlock)) {
@@ -311,7 +322,7 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
           // 找到在空块链中导致 currentDefBlock 的那个前驱块
           if (emptyBlockRedirectMap.count(incomingBlock) || incomingBlock == currentBlock) {
             // 递归追溯该传入值
-            return getUltimateSourceValue(phi->getIncomingValue(incomingBlock), incomingBlock);
+            return getUltimateSourceValue(phi->getValfromBlk(incomingBlock), incomingBlock);
           }
         }
       }
@@ -354,7 +365,7 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
 
             if (actualEmptyPredecessorOfS) {
               // 获取 Phi 节点原本从 actualEmptyPredecessorOfS 接收的值
-              Value *valueFromEmptyPredecessor = phiInst->getIncomingValue(actualEmptyPredecessorOfS);
+              Value *valueFromEmptyPredecessor = phiInst->getValfromBlk(actualEmptyPredecessorOfS);
 
               // 追溯这个值，找到它在非空块中的最终来源
               // currentBlock 是 P
@@ -364,12 +375,13 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
 
               // 替换 Phi 节点的传入块和传入值
               if (ultimateSourceValue) { // 确保成功追溯到有效来源
-                phiInst->replaceIncoming(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
+                // phiInst->replaceIncoming(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
+                phiInst->replaceIncomingBlock(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
               } else {
                 assert(false && "[DelEmptyBlock] Unable to trace a valid source for Phi instruction");
                 // 无法追溯到有效来源，这可能是个错误或特殊情况
                 // 此时可能需要移除该 Phi 项，或者插入一个 undef 值
-                phiInst->removeIncoming(actualEmptyPredecessorOfS);
+                phiInst->getValfromBlk(actualEmptyPredecessorOfS);
               }
             }
           } else {
@@ -421,7 +433,7 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
 
             if (actualEmptyPredecessorOfS) {
               // 获取 Phi 节点原本从 actualEmptyPredecessorOfS 接收的值
-              Value *valueFromEmptyPredecessor = phiInst->getIncomingValue(actualEmptyPredecessorOfS);
+              Value *valueFromEmptyPredecessor = phiInst->getValfromBlk(actualEmptyPredecessorOfS);
 
               // 追溯这个值，找到它在非空块中的最终来源
               // currentBlock 是 P
@@ -431,12 +443,13 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
 
               // 替换 Phi 节点的传入块和传入值
               if (ultimateSourceValue) { // 确保成功追溯到有效来源
-                phiInst->replaceIncoming(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
+                // phiInst->replaceIncoming(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
+                phiInst->replaceIncomingBlock(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
               } else {
                 assert(false && "[DelEmptyBlock] Unable to trace a valid source for Phi instruction");
                 // 无法追溯到有效来源，这可能是个错误或特殊情况
                 // 此时可能需要移除该 Phi 项，或者插入一个 undef 值
-                phiInst->removeIncoming(actualEmptyPredecessorOfS);
+                phiInst->removeIncomingBlock(actualEmptyPredecessorOfS);
               }
             }
           } else {
@@ -481,7 +494,7 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
 
             if (actualEmptyPredecessorOfS) {
               // 获取 Phi 节点原本从 actualEmptyPredecessorOfS 接收的值
-              Value *valueFromEmptyPredecessor = phiInst->getIncomingValue(actualEmptyPredecessorOfS);
+              Value *valueFromEmptyPredecessor = phiInst->getValfromBlk(actualEmptyPredecessorOfS);
 
               // 追溯这个值，找到它在非空块中的最终来源
               // currentBlock 是 P
@@ -491,12 +504,13 @@ bool SysYCFGOptUtils::SysYDelEmptyBlock(Function *func, IRBuilder *pBuilder) {
 
               // 替换 Phi 节点的传入块和传入值
               if (ultimateSourceValue) { // 确保成功追溯到有效来源
-                phiInst->replaceIncoming(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
+                // phiInst->replaceIncoming(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
+                phiInst->replaceIncomingBlock(actualEmptyPredecessorOfS, currentBlock, ultimateSourceValue);
               } else {
                 assert(false && "[DelEmptyBlock] Unable to trace a valid source for Phi instruction");
                 // 无法追溯到有效来源，这可能是个错误或特殊情况
                 // 此时可能需要移除该 Phi 项，或者插入一个 undef 值
-                phiInst->removeIncoming(actualEmptyPredecessorOfS);
+                phiInst->removeIncomingBlock(actualEmptyPredecessorOfS);
               }
             }
           } else {
@@ -647,7 +661,7 @@ bool SysYCFGOptUtils::SysYCondBr2Br(Function *func, IRBuilder *pBuilder) {
               break;
             }
             // 使用 delBlk 方法删除 basicblock.get() 对应的传入值
-            dynamic_cast<PhiInst *>(phiinst.get())->removeIncoming(basicblock.get());
+            dynamic_cast<PhiInst *>(phiinst.get())->removeIncomingBlock(basicblock.get());
           }
 
         } else { // cond为false或0
@@ -665,7 +679,7 @@ bool SysYCFGOptUtils::SysYCondBr2Br(Function *func, IRBuilder *pBuilder) {
               break;
             }
             // 使用 delBlk 方法删除 basicblock.get() 对应的传入值
-            dynamic_cast<PhiInst *>(phiinst.get())->removeIncoming(basicblock.get());
+            dynamic_cast<PhiInst *>(phiinst.get())->removeIncomingBlock(basicblock.get());
           }
         }
       }

src/midend/Pass/Optimize/TailCallOpt.cpp

@@ -0,0 +1,125 @@
+#include "TailCallOpt.h"
+#include "IR.h"
+#include "IRBuilder.h"
+#include "SysYIROptUtils.h"
+#include <vector>
+// #include <iostream>
+#include <algorithm>
+
+namespace sysy {
+
+void *TailCallOpt::ID = (void *)&TailCallOpt::ID;
+
+void TailCallOpt::getAnalysisUsage(std::set<void *> &analysisDependencies, std::set<void *> &analysisInvalidations) const {
+  analysisInvalidations.insert(&DominatorTreeAnalysisPass::ID);
+  analysisInvalidations.insert(&LoopAnalysisPass::ID);
+}
+
+bool TailCallOpt::runOnFunction(Function *F, AnalysisManager &AM) {
+  std::vector<CallInst *> tailCallInsts;
+  // 遍历函数的所有基本块
+  for (auto &bb_ptr : F->getBasicBlocks()) {
+    auto BB = bb_ptr.get();
+    if (BB->getInstructions().empty()) continue; // 跳过空基本块
+
+    auto term_iter = BB->terminator();
+    if (term_iter == BB->getInstructions().end()) continue; // 没有终结指令则跳过
+    auto term = (*term_iter).get();
+
+    if (!term || !term->isReturn()) continue; // 不是返回指令则跳过
+    auto retInst = static_cast<ReturnInst *>(term);
+
+    Instruction *prevInst = nullptr;
+    if (BB->getInstructions().size() > 1) {
+        auto it = term_iter;
+        --it; // 获取返回指令前的指令
+        prevInst = (*it).get();
+    }
+
+    if (!prevInst || !prevInst->isCall()) continue; // 前一条不是调用指令则跳过
+    auto callInst = static_cast<CallInst *>(prevInst);
+
+    // 检查是否为尾递归调用：被调用函数与当前函数相同且返回值与调用结果匹配
+    if (callInst->getCallee() == F) {
+  // 对于尾递归，返回值应为调用结果或为 void 类型
+        if (retInst->getReturnValue() == callInst || 
+            (retInst->getReturnValue() == nullptr && callInst->getType()->isVoid())) {
+            tailCallInsts.push_back(callInst);
+        }
+    }
+  }
+
+  if (tailCallInsts.empty()) {
+    return false;
+  }
+
+  // 创建一个新的入口基本块，作为循环的前置块
+  auto original_entry = F->getEntryBlock();
+  auto new_entry = F->addBasicBlock("tco.entry." + F->getName());
+  auto loop_header = F->addBasicBlock("tco.loop_header." + F->getName());
+  
+  // 将原入口块中的所有指令移动到循环头块
+  loop_header->getInstructions().splice(loop_header->end(), original_entry->getInstructions());
+  original_entry->setName("tco.pre_header");
+
+  // 为函数参数创建 phi 节点
+  builder->setPosition(loop_header, loop_header->begin());
+  std::vector<PhiInst *> phis;
+  auto original_args = F->getArguments();
+  for (auto &arg : original_args) {
+    auto phi = builder->createPhiInst(arg->getType(), {}, {}, "tco.phi."+arg->getName());
+    phis.push_back(phi);
+  }
+
+  // 用 phi 节点替换所有原始参数的使用
+  for (size_t i = 0; i < original_args.size(); ++i) {
+    original_args[i]->replaceAllUsesWith(phis[i]);
+  }
+
+  // 设置 phi 节点的输入值
+  for (size_t i = 0; i < phis.size(); ++i) {
+    phis[i]->addIncoming(original_args[i], new_entry);
+  }
+
+  // 连接各个基本块
+  builder->setPosition(original_entry, original_entry->end());
+  builder->createUncondBrInst(new_entry);
+  original_entry->addSuccessor(new_entry);
+  
+  builder->setPosition(new_entry, new_entry->end());
+  builder->createUncondBrInst(loop_header);
+  new_entry->addSuccessor(loop_header);
+  loop_header->addPredecessor(new_entry);
+
+  // 处理每一个尾递归调用
+  for (auto callInst : tailCallInsts) {
+    auto tail_call_block = callInst->getParent();
+    
+  // 收集尾递归调用的参数
+    auto args_range = callInst->getArguments();
+    std::vector<Value*> args;
+    std::transform(args_range.begin(), args_range.end(), std::back_inserter(args), 
+                   [](auto& use_ptr){ return use_ptr->getValue(); });
+
+  // 用新的参数值更新 phi 节点
+    for (size_t i = 0; i < phis.size(); ++i) {
+        phis[i]->addIncoming(args[i], tail_call_block);
+    }
+
+  // 移除原有的调用和返回指令
+    auto term_iter = tail_call_block->terminator();
+    SysYIROptUtils::usedelete(term_iter);
+    auto call_iter = tail_call_block->findInstIterator(callInst);
+    SysYIROptUtils::usedelete(call_iter);
+
+  // 添加跳转回循环头块的分支指令
+    builder->setPosition(tail_call_block, tail_call_block->end());
+    builder->createUncondBrInst(loop_header);
+    tail_call_block->addSuccessor(loop_header);
+    loop_header->addPredecessor(tail_call_block);
+  }
+
+  return true;
+}
+
+} // namespace sysy

src/midend/Pass/Pass.cpp

@@ -1,11 +1,25 @@
 #include "Dom.h"
 #include "Liveness.h"
+#include "Loop.h"
+#include "LoopCharacteristics.h"
+#include "AliasAnalysis.h"
+#include "CallGraphAnalysis.h"
+#include "SideEffectAnalysis.h"
 #include "SysYIRCFGOpt.h"
 #include "SysYIRPrinter.h"
 #include "DCE.h"
 #include "Mem2Reg.h"
 #include "Reg2Mem.h"
+#include "GVN.h"
 #include "SCCP.h"
+#include "BuildCFG.h"
+#include "LargeArrayToGlobal.h"
+#include "LoopNormalization.h"
+#include "LICM.h"
+#include "LoopStrengthReduction.h"
+#include "InductionVariableElimination.h"
+#include "GlobalStrengthReduction.h"
+#include "TailCallOpt.h"
 #include "Pass.h"
 #include <iostream>
 #include <queue>
@@ -35,10 +49,22 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
         3. 添加优化passid
     */
     // 注册分析遍
+    registerAnalysisPass<DominatorTreeAnalysisPass>();
+    registerAnalysisPass<LivenessAnalysisPass>();
     registerAnalysisPass<sysy::DominatorTreeAnalysisPass>();
     registerAnalysisPass<sysy::LivenessAnalysisPass>();
+    registerAnalysisPass<SysYAliasAnalysisPass>();           // 别名分析 (优先级高)
+    registerAnalysisPass<CallGraphAnalysisPass>();           // 调用图分析 (Module级别，独立分析)
+    registerAnalysisPass<SysYSideEffectAnalysisPass>();      // 副作用分析 (依赖别名分析和调用图)
+    registerAnalysisPass<LoopAnalysisPass>();
+    registerAnalysisPass<LoopCharacteristicsPass>();        // 循环特征分析依赖别名分析
 
     // 注册优化遍
+    registerOptimizationPass<BuildCFG>();
+    registerOptimizationPass<LargeArrayToGlobalPass>();
+
+    registerOptimizationPass<GVN>();
+    
     registerOptimizationPass<SysYDelInstAfterBrPass>();
     registerOptimizationPass<SysYDelNoPreBLockPass>();
     registerOptimizationPass<SysYBlockMergePass>();
@@ -49,7 +75,14 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
 
     registerOptimizationPass<DCE>();
     registerOptimizationPass<Mem2Reg>(builderIR);
+    registerOptimizationPass<LoopNormalizationPass>(builderIR);
+    registerOptimizationPass<LICM>(builderIR);
+    registerOptimizationPass<LoopStrengthReduction>(builderIR);
+    registerOptimizationPass<InductionVariableElimination>();
+
+    registerOptimizationPass<GlobalStrengthReduction>(builderIR);
     registerOptimizationPass<Reg2Mem>(builderIR);
+    registerOptimizationPass<TailCallOpt>(builderIR);
 
     registerOptimizationPass<SCCP>(builderIR);
 
@@ -58,6 +91,16 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
       if (DEBUG) std::cout << "Applying -O1 optimizations.\n";
       if (DEBUG) std::cout << "--- Running custom optimization sequence ---\n";
 
+      if(DEBUG) {
+        std::cout << "=== IR Before CFGOpt Optimizations ===\n";
+        printPasses();
+      }
+
+      this->clearPasses();
+      this->addPass(&BuildCFG::ID);
+      this->addPass(&LargeArrayToGlobalPass::ID);
+      this->run();
+
       this->clearPasses(); 
       this->addPass(&SysYDelInstAfterBrPass::ID);
       this->addPass(&SysYDelNoPreBLockPass::ID);
@@ -67,6 +110,10 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
       this->addPass(&SysYAddReturnPass::ID);
       this->run(); 
 
+      this->clearPasses();
+      this->addPass(&BuildCFG::ID);
+      this->run();
+
       if(DEBUG) {
         std::cout << "=== IR After CFGOpt Optimizations ===\n";
         printPasses();
@@ -90,6 +137,25 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
         printPasses();
       }
 
+      this->clearPasses();
+      this->addPass(&GVN::ID);
+      this->run();
+
+      this->clearPasses();
+      this->addPass(&TailCallOpt::ID);
+      this->run();
+
+      if(DEBUG) {
+        std::cout << "=== IR After TailCallOpt ===\n";
+        SysYPrinter printer(moduleIR);
+        printer.printIR();
+      }
+
+      if(DEBUG) {
+        std::cout << "=== IR After GVN Optimizations ===\n";
+        printPasses();
+      }
+
       this->clearPasses();
       this->addPass(&SCCP::ID);
       this->run();
@@ -99,6 +165,45 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
         printPasses();
       }
 
+      this->clearPasses();
+      this->addPass(&LoopNormalizationPass::ID);
+      this->addPass(&InductionVariableElimination::ID);
+      this->run();
+
+      if(DEBUG) {
+        std::cout << "=== IR After Loop Normalization, Induction Variable Elimination ===\n";
+        printPasses();
+      }
+      
+
+      this->clearPasses();
+      this->addPass(&LICM::ID);
+      this->run();
+
+      if(DEBUG) {
+        std::cout << "=== IR After LICM ===\n";
+        printPasses();
+      }
+      
+      this->clearPasses();
+      this->addPass(&LoopStrengthReduction::ID);
+      this->run();
+
+      if(DEBUG) {
+        std::cout << "=== IR After Loop Normalization, and Strength Reduction Optimizations ===\n";
+        printPasses();
+      }
+
+      // 全局强度削弱优化，包括代数优化和魔数除法
+      this->clearPasses();
+      this->addPass(&GlobalStrengthReduction::ID);
+      this->run();
+
+      if(DEBUG) {
+        std::cout << "=== IR After Global Strength Reduction Optimizations ===\n";
+        printPasses();
+      }
+
       this->clearPasses();
       this->addPass(&Reg2Mem::ID);
       this->run();
@@ -107,7 +212,9 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
         std::cout << "=== IR After Reg2Mem Optimizations ===\n";
         printPasses();
       }
-
+      this->clearPasses();
+      this->addPass(&BuildCFG::ID);
+      this->run();
       if (DEBUG) std::cout << "--- Custom optimization sequence finished ---\n";
     }
 
@@ -122,6 +229,7 @@ void PassManager::runOptimizationPipeline(Module* moduleIR, IRBuilder* builderIR
       SysYPrinter printer(moduleIR);
       printer.printIR();
     }
+    
 }
 
 void PassManager::clearPasses() {

src/midend/SysYIRGenerator.cpp

@@ -15,6 +15,139 @@
 using namespace std;
 namespace sysy {
 
+std::pair<long long, int> calculate_signed_magic(int d) {
+    if (d == 0) throw std::runtime_error("Division by zero");
+    if (d == 1 || d == -1) return {0, 0}; // Not used by strength reduction
+
+    int k = 0;
+    unsigned int ad = (d > 0) ? d : -d;
+    unsigned int temp = ad;
+    while (temp > 0) {
+        temp >>= 1;
+        k++;
+    }
+    if ((ad & (ad - 1)) == 0) { // if power of 2
+        k--;
+    }
+
+    unsigned __int128 m_val = 1;
+    m_val <<= (32 + k - 1);
+    unsigned __int128 m_prime = m_val / ad;
+    long long m = m_prime + 1;
+
+    return {m, k};
+}
+
+// 清除因函数调用而失效的表达式缓存（保守策略）
+void SysYIRGenerator::invalidateExpressionsOnCall() {
+  availableBinaryExpressions.clear();
+  availableUnaryExpressions.clear();
+  availableLoads.clear();
+  availableGEPs.clear();
+}
+
+// 在进入新的基本块时清空所有表达式缓存
+void SysYIRGenerator::enterNewBasicBlock() {
+  availableBinaryExpressions.clear();
+  availableUnaryExpressions.clear();
+  availableLoads.clear();
+  availableGEPs.clear();
+}
+
+// 清除因变量赋值而失效的表达式缓存
+// @param storedAddress: store 指令的目标地址 (例如 AllocaInst* 或 GEPInst*)
+void SysYIRGenerator::invalidateExpressionsOnStore(Value *storedAddress) {
+  // 遍历二元表达式缓存，移除受影响的条目
+  // 创建一个临时列表来存储要移除的键，避免在迭代时修改容器
+  std::vector<ExpKey> binaryKeysToRemove;
+  for (const auto &pair : availableBinaryExpressions) {
+    // 检查左操作数
+    // 如果左操作数是 LoadInst，并且它从 storedAddress 加载
+    if (auto loadInst = dynamic_cast<LoadInst *>(pair.first.left)) {
+      if (loadInst->getPointer() == storedAddress) {
+        binaryKeysToRemove.push_back(pair.first);
+        continue; // 这个表达式已标记为移除，跳到下一个
+      }
+    }
+    // 如果左操作数本身就是被存储的地址 (例如，将一个地址值直接作为操作数，虽然不常见)
+    if (pair.first.left == storedAddress) {
+      binaryKeysToRemove.push_back(pair.first);
+      continue;
+    }
+
+    // 检查右操作数，逻辑同左操作数
+    if (auto loadInst = dynamic_cast<LoadInst *>(pair.first.right)) {
+      if (loadInst->getPointer() == storedAddress) {
+        binaryKeysToRemove.push_back(pair.first);
+        continue;
+      }
+    }
+    if (pair.first.right == storedAddress) {
+      binaryKeysToRemove.push_back(pair.first);
+      continue;
+    }
+  }
+  // 实际移除条目
+  for (const auto &key : binaryKeysToRemove) {
+    availableBinaryExpressions.erase(key);
+  }
+
+  // 遍历一元表达式缓存，移除受影响的条目
+  std::vector<UnExpKey> unaryKeysToRemove;
+  for (const auto &pair : availableUnaryExpressions) {
+    // 检查操作数
+    if (auto loadInst = dynamic_cast<LoadInst *>(pair.first.operand)) {
+      if (loadInst->getPointer() == storedAddress) {
+        unaryKeysToRemove.push_back(pair.first);
+        continue;
+      }
+    }
+    if (pair.first.operand == storedAddress) {
+      unaryKeysToRemove.push_back(pair.first);
+      continue;
+    }
+  }
+  // 实际移除条目
+  for (const auto &key : unaryKeysToRemove) {
+    availableUnaryExpressions.erase(key);
+  }
+  availableLoads.erase(storedAddress);
+
+  std::vector<GEPKey> gepKeysToRemove;
+  for (const auto &pair : availableGEPs) {
+    // 检查 GEP 的基指针是否受存储影响
+    if (auto loadInst = dynamic_cast<LoadInst *>(pair.first.basePointer)) {
+        if (loadInst->getPointer() == storedAddress) {
+            gepKeysToRemove.push_back(pair.first);
+            continue; // 标记此GEP为移除，跳过后续检查
+        }
+    }
+    // 如果基指针本身就是存储的目标地址 (不常见，但可能)
+    if (pair.first.basePointer == storedAddress) {
+        gepKeysToRemove.push_back(pair.first);
+        continue;
+    }
+
+    // 检查 GEP 的每个索引是否受存储影响
+    for (const auto &indexVal : pair.first.indices) {
+        if (auto loadInst = dynamic_cast<LoadInst *>(indexVal)) {
+            if (loadInst->getPointer() == storedAddress) {
+                gepKeysToRemove.push_back(pair.first);
+                break; // 标记此GEP为移除，并跳出内部循环
+            }
+        }
+        // 如果索引本身就是存储的目标地址
+        if (indexVal == storedAddress) {
+            gepKeysToRemove.push_back(pair.first);
+            break;
+        }
+    }
+  }
+  // 实际移除条目
+  for (const auto &key : gepKeysToRemove) {
+    availableGEPs.erase(key);
+  }
+}
 
 // std::vector<Value*> BinaryValueStack;  ///< 用于存储value的栈
 // std::vector<int> BinaryOpStack;  ///< 用于存储二元表达式的操作符栈 
@@ -244,27 +377,37 @@ void SysYIRGenerator::compute() {
           }
         } else {
           // 否则，创建相应的IR指令
-          if (commonType == Type::getIntType()) {
-            switch (op) {
-            case BinaryOp::ADD: resultValue = builder.createAddInst(lhs, rhs); break;
-            case BinaryOp::SUB: resultValue = builder.createSubInst(lhs, rhs); break;
-            case BinaryOp::MUL: resultValue = builder.createMulInst(lhs, rhs); break;
-            case BinaryOp::DIV: resultValue = builder.createDivInst(lhs, rhs); break;
-            case BinaryOp::MOD: resultValue = builder.createRemInst(lhs, rhs); break;
-            }
-          } else if (commonType == Type::getFloatType()) {
-            switch (op) {
-            case BinaryOp::ADD: resultValue = builder.createFAddInst(lhs, rhs); break;
-            case BinaryOp::SUB: resultValue = builder.createFSubInst(lhs, rhs); break;
-            case BinaryOp::MUL: resultValue = builder.createFMulInst(lhs, rhs); break;
-            case BinaryOp::DIV: resultValue = builder.createFDivInst(lhs, rhs); break;
-            case BinaryOp::MOD:
-              std::cerr << "Error: Modulo operator not supported for float types." << std::endl;
+          ExpKey currentExpKey(static_cast<BinaryOp>(op), lhs, rhs);
+          auto it = availableBinaryExpressions.find(currentExpKey);
+
+          if (it != availableBinaryExpressions.end()) {
+            // 在缓存中找到，重用结果
+            resultValue = it->second;
+          } else {
+            if (commonType == Type::getIntType()) {
+              switch (op) {
+              case BinaryOp::ADD: resultValue = builder.createAddInst(lhs, rhs); break;
+              case BinaryOp::SUB: resultValue = builder.createSubInst(lhs, rhs); break;
+              case BinaryOp::MUL: resultValue = builder.createMulInst(lhs, rhs); break;
+              case BinaryOp::DIV: resultValue = builder.createDivInst(lhs, rhs); break;
+              case BinaryOp::MOD: resultValue = builder.createRemInst(lhs, rhs); break;
+              }
+            } else if (commonType == Type::getFloatType()) {
+              switch (op) {
+              case BinaryOp::ADD: resultValue = builder.createFAddInst(lhs, rhs); break;
+              case BinaryOp::SUB: resultValue = builder.createFSubInst(lhs, rhs); break;
+              case BinaryOp::MUL: resultValue = builder.createFMulInst(lhs, rhs); break;
+              case BinaryOp::DIV: resultValue = builder.createFDivInst(lhs, rhs); break;
+              case BinaryOp::MOD:
+                std::cerr << "Error: Modulo operator not supported for float types." << std::endl;
+                return;
+              }
+            } else {
+              std::cerr << "Error: Unsupported type for binary instruction." << std::endl;
               return;
             }
-          } else {
-            std::cerr << "Error: Unsupported type for binary instruction." << std::endl;
-            return;
+            // 将新创建的指令结果添加到缓存
+            availableBinaryExpressions[currentExpKey] = resultValue;
           }
         }
         break;
@@ -316,36 +459,45 @@ void SysYIRGenerator::compute() {
             return;
           }
         } else {
-          // 否则，创建相应的IR指令
-          switch (op) {
-          case BinaryOp::PLUS:
-            resultValue = operand; // 一元加指令通常直接返回操作数
-            break;
-          case BinaryOp::NEG: {
-            if (commonType == sysy::Type::getIntType()) {
-              resultValue = builder.createNegInst(operand);
-            } else if (commonType == sysy::Type::getFloatType()) {
-              resultValue = builder.createFNegInst(operand);
-            } else {
-              std::cerr << "Error: Negation not supported for operand type." << std::endl;
-              return;
+          // 否则，创建相应的IR指令 (在这里应用CSE)
+          UnExpKey currentUnExpKey(static_cast<BinaryOp>(op), operand);
+          auto it = availableUnaryExpressions.find(currentUnExpKey);
+          if (it != availableUnaryExpressions.end()) {
+            // 在缓存中找到，重用结果
+            resultValue = it->second;
+          } else {
+            switch (op) {
+              case BinaryOp::PLUS:
+                resultValue = operand; // 一元加指令通常直接返回操作数
+                break;
+              case BinaryOp::NEG: {
+                if (commonType == sysy::Type::getIntType()) {
+                  resultValue = builder.createNegInst(operand);
+                } else if (commonType == sysy::Type::getFloatType()) {
+                  resultValue = builder.createFNegInst(operand);
+                } else {
+                  std::cerr << "Error: Negation not supported for operand type." << std::endl;
+                  return;
+                }
+                break;
+              }
+              case BinaryOp::NOT:
+                // 逻辑非
+                if (commonType == sysy::Type::getIntType()) {
+                  resultValue = builder.createNotInst(operand);
+                } else if (commonType == sysy::Type::getFloatType()) {
+                  resultValue = builder.createFNotInst(operand);
+                } else {
+                  std::cerr << "Error: Logical NOT not supported for operand type." << std::endl;
+                  return;
+                }
+                break;
+              default:
+                std::cerr << "Error: Unknown unary operator for instructions: " << op << std::endl;
+                return;
             }
-            break;
-          }
-          case BinaryOp::NOT:
-            // 逻辑非
-            if (commonType == sysy::Type::getIntType()) {
-              resultValue = builder.createNotInst(operand);
-            } else if (commonType == sysy::Type::getFloatType()) {
-              resultValue = builder.createFNotInst(operand);
-            } else {
-              std::cerr << "Error: Logical NOT not supported for operand type." << std::endl;
-              return;
-            }
-            break;
-          default:
-            std::cerr << "Error: Unknown unary operator for instructions: " << op << std::endl;
-            return;
+            // 将新创建的指令结果添加到缓存
+            availableUnaryExpressions[currentUnExpKey] = resultValue;
           }
         }
         break;
@@ -487,7 +639,19 @@ Value* SysYIRGenerator::getGEPAddressInst(Value* basePointer, const std::vector<
     // `indices` 向量现在由调用方（如 visitLValue, visitVarDecl, visitAssignStmt）负责完整准备，
     // 包括是否需要添加初始的 `0` 索引。
     // 所以这里直接将其传递给 `builder.createGetElementPtrInst`。
-    return builder.createGetElementPtrInst(basePointer, indices);
+    GEPKey key = {basePointer, indices};
+
+    // 尝试从缓存中查找
+    auto it = availableGEPs.find(key);
+    if (it != availableGEPs.end()) {
+        return it->second; // 缓存命中，返回已有的 GEPInst*
+    }
+
+    // 缓存未命中，创建新的 GEPInst
+    Value* gepInst = builder.createGetElementPtrInst(basePointer, indices); // 假设 builder 提供了 createGEPInst 方法
+    availableGEPs[key] = gepInst; // 将新的 GEPInst* 加入缓存
+
+    return gepInst;
 }
 
 /*
@@ -586,7 +750,13 @@ std::any SysYIRGenerator::visitConstDecl(SysYParser::ConstDeclContext *ctx) {
 
     // 显式地为局部常量在栈上分配空间
     // alloca 的类型将是指针指向常量类型，例如 `int*` 或 `int[2][3]*`
+    // 将alloca全部集中到entry中
+    auto entry = builder.getBasicBlock()->getParent()->getEntryBlock();
+    auto it = builder.getPosition();
+    auto nowblk = builder.getBasicBlock();
+    builder.setPosition(entry, entry->terminator());
     AllocaInst *alloca = builder.createAllocaInst(Type::getPointerType(variableType), name);
+    builder.setPosition(nowblk, it);
 
     ArrayValueTree *root = std::any_cast<ArrayValueTree *>(constDef->constInitVal()->accept(this));
     ValueCounter values;
@@ -743,8 +913,12 @@ std::any SysYIRGenerator::visitVarDecl(SysYParser::VarDeclContext *ctx) {
 
     // 对于数组，alloca 的类型将是指针指向数组类型，例如 `int[2][3]*`
     // 对于标量，alloca 的类型将是指针指向标量类型，例如 `int*`
-    AllocaInst* alloca =
-        builder.createAllocaInst(Type::getPointerType(variableType), name);
+    auto entry = builder.getBasicBlock()->getParent()->getEntryBlock();
+    auto it = builder.getPosition();
+    auto nowblk = builder.getBasicBlock();
+    builder.setPosition(entry, entry->terminator());
+    AllocaInst *alloca = builder.createAllocaInst(Type::getPointerType(variableType), name);
+    builder.setPosition(nowblk, it);
 
     if (varDef->initVal() != nullptr) {
       ValueCounter values;
@@ -946,6 +1120,8 @@ std::any SysYIRGenerator::visitFuncType(SysYParser::FuncTypeContext *ctx) {
 std::any SysYIRGenerator::visitFuncDef(SysYParser::FuncDefContext *ctx){
   // 更新作用域
   module->enterNewScope();
+  // 清除CSE缓存
+  enterNewBasicBlock();
 
   auto name = ctx->Ident()->getText();
   std::vector<Type *> paramActualTypes;
@@ -1015,15 +1191,25 @@ std::any SysYIRGenerator::visitFuncDef(SysYParser::FuncDefContext *ctx){
   for(int i = 0; i < paramActualTypes.size(); ++i) {
     Argument* arg = new Argument(paramActualTypes[i], function, i, paramNames[i]);
     function->insertArgument(arg);
+  }
 
+  // 先将所有参数名字注册到符号表中，确保alloca不会使用相同的名字
+  for (int i = 0; i < paramNames.size(); ++i) {
+    // 预先注册参数名字，这样addVariable就会使用不同的后缀
+    module->registerParameterName(paramNames[i]);
   }
 
   auto funcArgs = function->getArguments();
   std::vector<AllocaInst *> allocas;
   for (int i = 0; i < paramActualTypes.size(); ++i) {
-    AllocaInst *alloca = builder.createAllocaInst(Type::getPointerType(paramActualTypes[i]), paramNames[i]);
+    // 使用函数特定的前缀来确保参数alloca名字唯一
+    std::string allocaName = name + "_param_" + paramNames[i];
+    AllocaInst *alloca = builder.createAllocaInst(Type::getPointerType(paramActualTypes[i]), allocaName);
+    // 直接设置唯一名字，不依赖addVariable的命名逻辑
+    alloca->setName(allocaName);
     allocas.push_back(alloca);
-    module->addVariable(paramNames[i], alloca);
+    // 直接添加到符号表，使用原参数名作为查找键
+    module->addVariableDirectly(paramNames[i], alloca);
   }
 
   for(int i = 0; i < paramActualTypes.size(); ++i) {
@@ -1037,6 +1223,7 @@ std::any SysYIRGenerator::visitFuncDef(SysYParser::FuncDefContext *ctx){
   
   // 从 entryBB 无条件跳转到 funcBodyEntry
   builder.createUncondBrInst(funcBodyEntry);
+  BasicBlock::conectBlocks(entry, funcBodyEntry); // 连接 entryBB 和 funcBodyEntry
   builder.setPosition(funcBodyEntry,funcBodyEntry->end()); // 将插入点设置到 funcBodyEntry
 
   for (auto item : ctx->blockStmt()->blockItem()) {
@@ -1091,6 +1278,45 @@ std::any SysYIRGenerator::visitAssignStmt(SysYParser::AssignStmtContext *ctx) {
     if (dynamic_cast<AllocaInst*>(variable) || dynamic_cast<GlobalValue*>(variable)) {
         LValue = variable;
     }
+    
+    // 标量变量的类型推断
+    Type* LType = builder.getIndexedType(variable->getType(), indices);
+    
+    Value* RValue = computeExp(ctx->exp(), LType); // 右值计算
+    Type* RType = RValue->getType();
+
+    // TODO:computeExp处理了类型转换，可以考虑删除判断逻辑
+    if (LType != RType) {
+      ConstantValue *constValue = dynamic_cast<ConstantValue *>(RValue);
+      if (constValue != nullptr) {
+        if (LType == Type::getFloatType()) {
+          if(dynamic_cast<ConstantInteger *>(constValue)) {
+            // 如果是整型常量，转换为浮点型
+            RValue = ConstantFloating::get(static_cast<float>(constValue->getInt()));
+          } else if (dynamic_cast<ConstantFloating *>(constValue)) {
+            // 如果是浮点型常量，直接使用
+            RValue = ConstantFloating::get(static_cast<float>(constValue->getFloat()));
+          }
+        } else { // 假设如果不是浮点型，就是整型
+          if(dynamic_cast<ConstantFloating *>(constValue)) {
+            // 如果是浮点型常量，转换为整型
+            RValue = ConstantInteger::get(static_cast<int>(constValue->getFloat()));
+          } else if (dynamic_cast<ConstantInteger *>(constValue)) {
+            // 如果是整型常量，直接使用
+            RValue = ConstantInteger::get(static_cast<int>(constValue->getInt()));
+          }
+        }
+      } else {
+        if (LType == Type::getFloatType() && RType != Type::getFloatType()) {
+          RValue = builder.createItoFInst(RValue);
+        } else if (LType != Type::getFloatType() && RType == Type::getFloatType()) {
+          RValue = builder.createFtoIInst(RValue);
+        }
+        // 如果两者都是同一类型，就不需要转换
+      }
+    }
+    
+    builder.createStoreInst(RValue, LValue);
   }
   else {
     // 对于数组或多维数组的左值处理
@@ -1100,7 +1326,16 @@ std::any SysYIRGenerator::visitAssignStmt(SysYParser::AssignStmtContext *ctx) {
     if (AllocaInst *alloc = dynamic_cast<AllocaInst *>(variable)) {
       Type* allocatedType = alloc->getType()->as<PointerType>()->getBaseType(); 
       if (allocatedType->isPointer()) {
-        gepBasePointer = builder.createLoadInst(alloc); 
+        // 尝试从缓存中获取 builder.createLoadInst(alloc) 的结果
+        auto it = availableLoads.find(alloc);
+        if (it != availableLoads.end()) {
+            gepBasePointer = it->second; // 缓存命中，重用
+        } else {
+            gepBasePointer = builder.createLoadInst(alloc); // 缓存未命中，创建新的 LoadInst
+            availableLoads[alloc] = gepBasePointer; // 将结果加入缓存
+        }
+        // --- CSE 结束 ---
+        // gepBasePointer = builder.createLoadInst(alloc);
         gepIndices = indices; 
       } else {
         gepBasePointer = alloc;
@@ -1119,52 +1354,48 @@ std::any SysYIRGenerator::visitAssignStmt(SysYParser::AssignStmtContext *ctx) {
     } 
     // 左值为地址
     LValue = getGEPAddressInst(gepBasePointer, gepIndices);
-  }
+    
+    // 数组变量的类型推断，使用gepIndices和gepBasePointer的类型
+    Type* LType = builder.getIndexedType(gepBasePointer->getType(), gepIndices);
+    
+    Value* RValue = computeExp(ctx->exp(), LType); // 右值计算
+    Type* RType = RValue->getType();
 
-  // Value* RValue = std::any_cast<Value *>(visitExp(ctx->exp())); // 右值
-
-  // 先推断 LValue 的类型
-  // 如果 LValue 是指向数组的指针，则需要根据 indices 获取正确的类型
-  // 如果 LValue 是标量，则直接使用其类型
-  // 注意：LValue 的类型可能是指向数组的指针 (e.g., int(*)[3]) 或者指向标量的指针 (e.g., int*) 也能推断
-  Type* LType = builder.getIndexedType(variable->getType(), indices);
-
-  Value* RValue = computeExp(ctx->exp(), LType); // 右值计算
-  Type* RType = RValue->getType();
-
-  // TODO:computeExp处理了类型转换，可以考虑删除判断逻辑
-  if (LType != RType) {
-    ConstantValue *constValue = dynamic_cast<ConstantValue *>(RValue);
-    if (constValue != nullptr) {
-      if (LType == Type::getFloatType()) {
-        if(dynamic_cast<ConstantInteger *>(constValue)) {
-          // 如果是整型常量，转换为浮点型
-          RValue = ConstantFloating::get(static_cast<float>(constValue->getInt()));
-        } else if (dynamic_cast<ConstantFloating *>(constValue)) {
-          // 如果是浮点型常量，直接使用
-          RValue = ConstantFloating::get(static_cast<float>(constValue->getFloat()));
+    // TODO:computeExp处理了类型转换，可以考虑删除判断逻辑
+    if (LType != RType) {
+      ConstantValue *constValue = dynamic_cast<ConstantValue *>(RValue);
+      if (constValue != nullptr) {
+        if (LType == Type::getFloatType()) {
+          if(dynamic_cast<ConstantInteger *>(constValue)) {
+            // 如果是整型常量，转换为浮点型
+            RValue = ConstantFloating::get(static_cast<float>(constValue->getInt()));
+          } else if (dynamic_cast<ConstantFloating *>(constValue)) {
+            // 如果是浮点型常量，直接使用
+            RValue = ConstantFloating::get(static_cast<float>(constValue->getFloat()));
+          }
+        } else { // 假设如果不是浮点型，就是整型
+          if(dynamic_cast<ConstantFloating *>(constValue)) {
+            // 如果是浮点型常量，转换为整型
+            RValue = ConstantInteger::get(static_cast<int>(constValue->getFloat()));
+          } else if (dynamic_cast<ConstantInteger *>(constValue)) {
+            // 如果是整型常量，直接使用
+            RValue = ConstantInteger::get(static_cast<int>(constValue->getInt()));
+          }
         }
-      } else { // 假设如果不是浮点型，就是整型
-        if(dynamic_cast<ConstantFloating *>(constValue)) {
-          // 如果是浮点型常量，转换为整型
-          RValue = ConstantInteger::get(static_cast<int>(constValue->getFloat()));
-        } else if (dynamic_cast<ConstantInteger *>(constValue)) {
-          // 如果是整型常量，直接使用
-        RValue = ConstantInteger::get(static_cast<int>(constValue->getInt()));
-
+      } else {
+        if (LType == Type::getFloatType() && RType != Type::getFloatType()) {
+          RValue = builder.createItoFInst(RValue);
+        } else if (LType != Type::getFloatType() && RType == Type::getFloatType()) {
+          RValue = builder.createFtoIInst(RValue);
         }
-      }
-    } else {
-      if (LType == Type::getFloatType()) {
-        RValue = builder.createItoFInst(RValue);
-      } else { // 假设如果不是浮点型，就是整型
-        RValue = builder.createFtoIInst(RValue);
+        // 如果两者都是同一类型，就不需要转换
       }
     }
+    
+    builder.createStoreInst(RValue, LValue);
   }
 
-  builder.createStoreInst(RValue, LValue);
-
+  invalidateExpressionsOnStore(LValue);
   return std::any();
 }
 
@@ -1201,7 +1432,9 @@ std::any SysYIRGenerator::visitIfStmt(SysYParser::IfStmtContext *ctx) {
     labelstring.str("");
     function->addBasicBlock(thenBlock);
     builder.setPosition(thenBlock, thenBlock->end());
-
+    // CSE清除缓存
+    enterNewBasicBlock();
+    
     auto block = dynamic_cast<SysYParser::BlockStmtContext *>(ctx->stmt(0));
     // 如果是块语句，直接访问
     // 否则访问语句
@@ -1220,7 +1453,9 @@ std::any SysYIRGenerator::visitIfStmt(SysYParser::IfStmtContext *ctx) {
     labelstring.str("");
     function->addBasicBlock(elseBlock);
     builder.setPosition(elseBlock, elseBlock->end());
-
+    // CSE清除缓存
+    enterNewBasicBlock();
+    
     block = dynamic_cast<SysYParser::BlockStmtContext *>(ctx->stmt(1));
     if (block != nullptr) {
       visitBlockStmt(block);
@@ -1237,7 +1472,9 @@ std::any SysYIRGenerator::visitIfStmt(SysYParser::IfStmtContext *ctx) {
     labelstring.str("");
     function->addBasicBlock(exitBlock);
     builder.setPosition(exitBlock, exitBlock->end());
-
+    // CSE清除缓存
+    enterNewBasicBlock();
+    
   } else {
     builder.pushTrueBlock(thenBlock);
     builder.pushFalseBlock(exitBlock);
@@ -1250,7 +1487,9 @@ std::any SysYIRGenerator::visitIfStmt(SysYParser::IfStmtContext *ctx) {
     labelstring.str("");
     function->addBasicBlock(thenBlock);
     builder.setPosition(thenBlock, thenBlock->end());
-
+    // CSE清除缓存
+    enterNewBasicBlock();
+    
     auto block = dynamic_cast<SysYParser::BlockStmtContext *>(ctx->stmt(0));
     if (block != nullptr) {
       visitBlockStmt(block);
@@ -1267,6 +1506,9 @@ std::any SysYIRGenerator::visitIfStmt(SysYParser::IfStmtContext *ctx) {
     labelstring.str("");
     function->addBasicBlock(exitBlock);
     builder.setPosition(exitBlock, exitBlock->end());
+    // CSE清除缓存
+    enterNewBasicBlock();
+    
   }
   return std::any();
 }
@@ -1284,7 +1526,9 @@ std::any SysYIRGenerator::visitWhileStmt(SysYParser::WhileStmtContext *ctx) {
   builder.createUncondBrInst(headBlock);
   BasicBlock::conectBlocks(curBlock, headBlock);
   builder.setPosition(headBlock, headBlock->end());
-
+  // CSE清除缓存
+  enterNewBasicBlock();
+    
   BasicBlock* bodyBlock = new BasicBlock(function);
   BasicBlock* exitBlock = new BasicBlock(function);
 
@@ -1300,6 +1544,8 @@ std::any SysYIRGenerator::visitWhileStmt(SysYParser::WhileStmtContext *ctx) {
   labelstring.str("");
   function->addBasicBlock(bodyBlock);
   builder.setPosition(bodyBlock, bodyBlock->end());
+  // CSE清除缓存
+  enterNewBasicBlock();
 
   builder.pushBreakBlock(exitBlock);
   builder.pushContinueBlock(headBlock);
@@ -1315,7 +1561,7 @@ std::any SysYIRGenerator::visitWhileStmt(SysYParser::WhileStmtContext *ctx) {
   }
 
   builder.createUncondBrInst(headBlock);
-  BasicBlock::conectBlocks(builder.getBasicBlock(), exitBlock);
+  BasicBlock::conectBlocks(builder.getBasicBlock(), headBlock);
   builder.popBreakBlock();
   builder.popContinueBlock();
 
@@ -1324,7 +1570,9 @@ std::any SysYIRGenerator::visitWhileStmt(SysYParser::WhileStmtContext *ctx) {
   labelstring.str("");
   function->addBasicBlock(exitBlock);
   builder.setPosition(exitBlock, exitBlock->end());
-  
+    // CSE清除缓存
+  enterNewBasicBlock();
+
   return std::any();
 }
 
@@ -1430,90 +1678,101 @@ std::any SysYIRGenerator::visitLValue(SysYParser::LValueContext *ctx) {
         break;
       }
     }
-    if (allIndicesConstant) {
+    // 如果是常量变量且所有索引都是常量，并且不是数组名单独出现的情况
+    if (allIndicesConstant && !dims.empty()) {
       // 如果是常量变量且所有索引都是常量，直接通过 getByIndices 获取编译时值
       // 这个方法会根据索引深度返回最终的标量值或指向子数组的指针 (作为 ConstantValue/Variable)
       return constVar->getByIndices(dims);
     }
+    // 如果dims为空，检查是否是常量标量
+    if (dims.empty() && declaredNumDims == 0) {
+      // 常量标量，直接返回其值
+      // 默认传入空索引列表，表示访问标量本身
+      return constVar->getByIndices(dims);
+    }
+    // 如果dims为空但不是标量（数组名单独出现），需要走GEP路径来实现数组到指针的退化
   }
 
   // 3. 处理可变变量 (AllocaInst/GlobalValue) 或带非常量索引的常量变量
   // 这里区分标量访问和数组元素/子数组访问
-
+  Value *targetAddress = nullptr;
   // 检查是否是访问标量变量本身（没有索引，且声明维度为0）
   if (dims.empty() && declaredNumDims == 0) {
-    // 对于标量变量，直接加载其值。
-    // variable 本身就是指向标量的指针 (e.g., int* %a)
     if (dynamic_cast<AllocaInst*>(variable) || dynamic_cast<GlobalValue*>(variable)) {
-        value = builder.createLoadInst(variable);
-    } else {
-        // 如果走到这里且不是AllocaInst/GlobalValue，但dims为空且declaredNumDims为0，
-        // 且又不是ConstantVariable (前面已处理)，则可能是错误情况。
+        targetAddress = variable;
+    } 
+    else {
         assert(false && "Unhandled scalar variable type in LValue access.");
         return static_cast<Value*>(nullptr);
     }
   } else {
-    // 访问数组元素或子数组（有索引，或变量本身是数组/多维指针）
     Value* gepBasePointer = nullptr;
-    std::vector<Value*> gepIndices; // 准备传递给 getGEPAddressInst 的索引列表
-    // GEP 的基指针就是变量本身（它是一个指向内存的指针）
+    std::vector<Value*> gepIndices;
     if (AllocaInst *alloc = dynamic_cast<AllocaInst *>(variable)) {
-      // 情况 A: 局部变量 (AllocaInst)
-      // 获取 AllocaInst 分配的内存的实际类型。
-      // 例如：对于 `int b[10][20];`，`allocatedType` 是 `[10 x [20 x i32]]`。
-      // 对于 `int b[][20]` 的函数参数，其 AllocaInst 存储的是一个指针，
-      // 此时 `allocatedType` 是 `[20 x i32]*`。
       Type* allocatedType = alloc->getType()->as<PointerType>()->getBaseType(); 
-
       if (allocatedType->isPointer()) {
-        // 如果 AllocaInst 分配的是一个指针类型 (例如，用于存储函数参数的指针，如 int b[][20] 中的 b)
-        // 即 `allocatedType` 是一个指向数组指针的指针 (e.g., [20 x i32]**)
-        // 那么 GEP 的基指针是加载这个指针变量的值。
-        gepBasePointer = builder.createLoadInst(alloc); // 加载出实际的指针值 (e.g., [20 x i32]*)
-        // 对于这种参数指针，用户提供的索引直接作用于它。不需要额外的 0。
+        gepBasePointer = builder.createLoadInst(alloc);
         gepIndices = dims; 
       } else {
-        // 如果 AllocaInst 分配的是实际的数组数据 (例如，int b[10][20] 中的 b)
-        // 那么 AllocaInst 本身就是 GEP 的基指针。
-        // 这里的 `alloc` 是指向数组的指针 (e.g., [10 x [20 x i32]]*)
-        gepBasePointer = alloc; // 类型是 [10 x [20 x i32]]*
-        // 对于这种完整的数组分配，GEP 的第一个索引必须是 0，用于“步过”整个数组。
+        gepBasePointer = alloc;
         gepIndices.push_back(ConstantInteger::get(0));
-        gepIndices.insert(gepIndices.end(), dims.begin(), dims.end());
+        if (dims.empty() && declaredNumDims > 0) {
+          // 数组名单独出现（没有索引）：在SysY中，多维数组名应该退化为指向第一行的指针
+          // 对于二维数组 T[M][N]，退化为 T(*)[N]，需要GEP: getelementptr T[M][N], T[M][N]* ptr, i32 0, i32 0
+          // 第一个i32 0: 选择数组本身，第二个i32 0: 选择第0行
+          // 结果类型: T[N]*
+          gepIndices.push_back(ConstantInteger::get(0));
+        } else {
+          // 正常的数组元素访问
+          gepIndices.insert(gepIndices.end(), dims.begin(), dims.end());
+        }
       }
     } else if (GlobalValue *glob = dynamic_cast<GlobalValue *>(variable)) {
-      // 情况 B: 全局变量 (GlobalValue)
-      // GlobalValue 总是指向全局数据的指针。
-      gepBasePointer = glob; // 类型是 [61 x [67 x i32]]*
-      // 对于全局数组，GEP 的第一个索引必须是 0，用于“步过”整个数组。
+      gepBasePointer = glob;
       gepIndices.push_back(ConstantInteger::get(0));
-      gepIndices.insert(gepIndices.end(), dims.begin(), dims.end());
+      if (dims.empty() && declaredNumDims > 0) {
+        // 全局数组名单独出现（没有索引）：应该退化为指向第一行的指针
+        // 需要添加一个额外的i32 0索引
+        gepIndices.push_back(ConstantInteger::get(0));
+      } else {
+        // 正常的数组元素访问
+        gepIndices.insert(gepIndices.end(), dims.begin(), dims.end());
+      }
     } else if (ConstantVariable *constV = dynamic_cast<ConstantVariable *>(variable)) {
-      // 情况 C: 常量变量 (ConstantVariable)，如果它代表全局数组常量
-      // 假设 ConstantVariable 可以直接作为 GEP 的基指针。
       gepBasePointer = constV;
-      // 对于常量数组，也需要 0 索引来“步过”整个数组。
-      // 这里可以进一步检查 constV->getType()->as<PointerType>()->getBaseType()->isArray()
-      // 但为了简洁，假设所有 ConstantVariable 作为 GEP 基指针时都需要此 0。
       gepIndices.push_back(ConstantInteger::get(0));
-      gepIndices.insert(gepIndices.end(), dims.begin(), dims.end());
+      if (dims.empty() && declaredNumDims > 0) {
+        // 常量数组名单独出现（没有索引）：应该退化为指向第一行的指针
+        // 需要添加一个额外的i32 0索引
+        gepIndices.push_back(ConstantInteger::get(0));
+      } else {
+        // 正常的数组元素访问
+        gepIndices.insert(gepIndices.end(), dims.begin(), dims.end());
+      }
     } else {
       assert(false && "LValue variable type not supported for GEP base pointer.");
       return static_cast<Value *>(nullptr);
     }
 
-    // 现在调用 getGEPAddressInst，传入正确准备的基指针和索引列表
-    Value *targetAddress = getGEPAddressInst(gepBasePointer, gepIndices);
+    targetAddress = getGEPAddressInst(gepBasePointer, gepIndices);
 
-    // 如果提供的索引数量少于声明的维度数量，则表示访问的是子数组，返回其地址
-    if (dims.size() < declaredNumDims) {
-        value = targetAddress;
+  }
+
+  // 如果提供的索引数量少于声明的维度数量，则表示访问的是子数组，返回其地址 (无需加载)
+  if (dims.size() < declaredNumDims) {
+      value = targetAddress;
+  } else {
+    // value = builder.createLoadInst(targetAddress);
+    auto it = availableLoads.find(targetAddress);
+    if (it != availableLoads.end()) {
+      value = it->second; // 缓存命中，重用已有的 LoadInst 结果
     } else {
-        // 否则，表示访问的是最终的标量元素，加载其值
-        // 假设 createLoadInst 接受 Value* pointer
-        value = builder.createLoadInst(targetAddress);
+      // 缓存未命中，创建新的 LoadInst
+      value = builder.createLoadInst(targetAddress);
+      availableLoads[targetAddress] = value; // 将新的 LoadInst 结果加入缓存
     }
   }
+
   return value;
 }
 
@@ -1571,10 +1830,10 @@ std::any SysYIRGenerator::visitCall(SysYParser::CallContext *ctx) {
 
     // 获取形参列表。`getArguments()` 返回的是 `Argument*` 的集合，
     // 每个 `Argument` 代表一个函数形参，其 `getType()` 就是指向形参的类型的指针类型。
-    auto formalParams = function->getArguments();
+    const auto& formalParams = function->getArguments();
 
     // 检查实参和形参数量是否匹配。
-    if (args.size() != formalParams.size()) {
+    if (args.size() != function->getNumArguments()) {
       std::cerr << "Error: Function call argument count mismatch for function '" << funcName << "'." << std::endl;
       assert(false && "Function call argument count mismatch!");
     }
@@ -1606,15 +1865,27 @@ std::any SysYIRGenerator::visitCall(SysYParser::CallContext *ctx) {
           } else if (formalParamExpectedValueType->isFloat() && actualArgType->isInt()) {
             args[i] = builder.createItoFInst(args[i]);
           }
-          // 2. 指针类型转换 (例如数组退化：`[N x T]*` 到 `T*`，或兼容指针类型之间) TODO：不清楚有没有这种样例
+          // 2. 指针类型转换 (例如数组退化：`[N x T]*` 到 `T*`，或兼容指针类型之间)
           // 这种情况常见于数组参数，实参可能是一个更具体的数组指针类型，
-          // 而形参是其退化后的基础指针类型。LLVM 的 `bitcast` 指令可以用于
-          // 在相同大小的指针类型之间进行转换，这对于数组退化至关重要。
-          // else if (formalParamType->isPointer() && actualArgType->isPointer()) {
-            // 检查指针基类型是否兼容，或者是否是数组退化导致的类型不同。
-            // 使用 bitcast，
-            // args[i] = builder.createBitCastInst(args[i], formalParamType);
-          // }
+          // 而形参是其退化后的基础指针类型。
+          else if (formalParamExpectedValueType->isPointer() && actualArgType->isPointer()) {
+            // 检查是否是数组指针到元素指针的decay
+            // 例如：[N x T]* -> T*
+            auto formalPtrType = formalParamExpectedValueType->as<PointerType>();
+            auto actualPtrType = actualArgType->as<PointerType>();
+            
+            if (formalPtrType && actualPtrType && actualPtrType->getBaseType()->isArray()) {
+              auto actualArrayType = actualPtrType->getBaseType()->as<ArrayType>();
+              if (actualArrayType && 
+                  formalPtrType->getBaseType() == actualArrayType->getElementType()) {
+                // 这是数组decay的情况，添加GEP来获取数组的第一个元素
+                std::vector<Value*> indices;
+                indices.push_back(ConstantInteger::get(0)); // 第一个索引：解引用指针
+                indices.push_back(ConstantInteger::get(0)); // 第二个索引：获取数组第一个元素
+                args[i] = getGEPAddressInst(args[i], indices);
+              }
+            }
+          }
           // 3. 其他未预期的类型不匹配
           // 如果代码执行到这里，说明存在编译器前端未处理的类型不兼容或错误。
           else {
@@ -1633,6 +1904,7 @@ std::any SysYIRGenerator::visitUnaryExp(SysYParser::UnaryExpContext *ctx) {
     visitPrimaryExp(ctx->primaryExp());
   } else if (ctx->call() != nullptr) {
     BinaryExpStack.push_back(std::any_cast<Value *>(visitCall(ctx->call())));BinaryExpLenStack.back()++;
+    invalidateExpressionsOnCall();
   } else if (ctx->unaryOp() != nullptr) {
       // 遇到一元操作符，将其压入 BinaryExpStack
       auto opNode = dynamic_cast<antlr4::tree::TerminalNode*>(ctx->unaryOp()->children[0]);
@@ -1997,15 +2269,23 @@ void Utils::createExternalFunction(
     const std::vector<std::string> &paramNames,
     const std::vector<std::vector<Value *>> &paramDims, Type *returnType,
     const std::string &funcName, Module *pModule, IRBuilder *pBuilder) {
-  auto funcType = Type::getFunctionType(returnType, paramTypes);
+  // 根据paramDims调整参数类型，数组参数需要转换为指针类型
+  std::vector<Type *> adjustedParamTypes = paramTypes;
+  for (int i = 0; i < paramTypes.size() && i < paramDims.size(); ++i) {
+    if (!paramDims[i].empty()) {
+      // 如果参数有维度信息，说明是数组参数，转换为指针类型
+      adjustedParamTypes[i] = Type::getPointerType(paramTypes[i]);
+    }
+  }
+  auto funcType = Type::getFunctionType(returnType, adjustedParamTypes);
   auto function = pModule->createExternalFunction(funcName, funcType);
   auto entry = function->getEntryBlock();
   pBuilder->setPosition(entry, entry->end());
 
   for (int i = 0; i < paramTypes.size(); ++i) {
-    auto arg = new Argument(paramTypes[i], function, i, paramNames[i]);
+    auto arg = new Argument(adjustedParamTypes[i], function, i, paramNames[i]);
     auto alloca = pBuilder->createAllocaInst(
-        Type::getPointerType(paramTypes[i]), paramNames[i]);
+        Type::getPointerType(adjustedParamTypes[i]), paramNames[i]);
     function->insertArgument(arg);
     auto store = pBuilder->createStoreInst(arg, alloca);
     pModule->addVariable(paramNames[i], alloca);

src/midend/SysYIRPrinter.cpp

@@ -240,6 +240,10 @@ void SysYPrinter::printInst(Instruction *pInst) {
     case Kind::kMul:
     case Kind::kDiv:
     case Kind::kRem:
+    case Kind::kSrl:
+    case Kind::kSll:
+    case Kind::kSra:
+    case Kind::kMulh:
     case Kind::kFAdd:
     case Kind::kFSub:
     case Kind::kFMul:
@@ -272,6 +276,10 @@ void SysYPrinter::printInst(Instruction *pInst) {
         case Kind::kMul: std::cout << "mul"; break;
         case Kind::kDiv: std::cout << "sdiv"; break;
         case Kind::kRem: std::cout << "srem"; break;
+        case Kind::kSrl: std::cout << "lshr"; break;
+        case Kind::kSll: std::cout << "shl"; break;
+        case Kind::kSra: std::cout << "ashr"; break;
+        case Kind::kMulh: std::cout << "mulh"; break;
         case Kind::kFAdd: std::cout << "fadd"; break;
         case Kind::kFSub: std::cout << "fsub"; break;
         case Kind::kFMul: std::cout << "fmul"; break;
@@ -295,7 +303,12 @@ void SysYPrinter::printInst(Instruction *pInst) {
       
       // Types and operands
       std::cout << " ";
-      printType(binInst->getType());
+      // For comparison operations, print operand types instead of result type
+      if (pInst->getKind() >= Kind::kICmpEQ && pInst->getKind() <= Kind::kFCmpGE) {
+        printType(binInst->getLhs()->getType());
+      } else {
+        printType(binInst->getType());
+      }
       std::cout << " ";
       printValue(binInst->getLhs());
       std::cout << ", ";
@@ -508,9 +521,9 @@ void SysYPrinter::printInst(Instruction *pInst) {
         if (!firstPair) std::cout << ", ";
         firstPair = false;
         std::cout << "[ ";
-        printValue(phiInst->getValue(i));
+        printValue(phiInst->getIncomingValue(i));
         std::cout << ", %";
-        printBlock(phiInst->getBlock(i));
+        printBlock(phiInst->getIncomingBlock(i));
         std::cout << " ]";
       }
       std::cout << std::endl;

src/sysyc.cpp

@@ -35,7 +35,7 @@ void usage(int code) {
                     "Supported options:\n"
                     "  -h \tprint help message and exit\n"
                     "  -f \tpretty-format the input file\n"
-                    "  -s {ast,ir,asm,llvmir,asmd,ird}\tstop after generating AST/IR/Assembly\n"
+                    "  -s {ast,ir,asm,asmd,ird}\tstop after generating AST/IR/Assembly\n"
                     "  -S \tcompile to assembly (.s file)\n"
                     "  -o <file>\tplace the output into <file>\n"
                     "  -O<level>\tenable optimization at <level> (e.g., -O0, -O1)\n";
@@ -110,6 +110,7 @@ int main(int argc, char **argv) {
   // 如果指定停止在 AST 阶段，则打印并退出
   if (argStopAfter == "ast") {
     cout << moduleAST->toStringTree(true) << '\n';
+    sysy::cleanupIRPools(); // 清理内存池
     return EXIT_SUCCESS;
   }
 
@@ -132,7 +133,7 @@ int main(int argc, char **argv) {
   
   if (DEBUG) {
     cout << "=== Init IR ===\n";
-    SysYPrinter(moduleIR).printIR(); // 临时打印器用于调试
+    moduleIR->print(cout); // 使用新实现的print方法直接打印IR
   }
 
   // 创建 Pass 管理器并运行优化管道
@@ -144,10 +145,26 @@ int main(int argc, char **argv) {
   // a) 如果指定停止在 IR 阶段，则打印最终 IR 并退出
   if (argStopAfter == "ir" || argStopAfter == "ird") {
     // 打印最终 IR
-    cout << "=== Final IR ===\n";
-    SysYPrinter printer(moduleIR); // 在这里创建打印器，因为可能之前调试时用过临时打印器
-    printer.printIR();
+    if (DEBUG) cerr << "=== Final IR ===\n";
+    if (!argOutputFilename.empty()) {
+      // 输出到指定文件
+      ofstream fout(argOutputFilename);
+      if (not fout.is_open()) {
+        cerr << "Failed to open output file: " << argOutputFilename << endl;
+        moduleIR->cleanup(); // 清理模块
+        sysy::cleanupIRPools(); // 清理内存池
+        return EXIT_FAILURE;
+      }
+      moduleIR->print(fout);
+      fout.close();
+    } else {
+      // 输出到标准输出
+      moduleIR->print(cout);
+    }
+    moduleIR->cleanup(); // 清理模块
+    sysy::cleanupIRPools(); // 清理内存池
     return EXIT_SUCCESS;
+
   }
 
   // b) 如果未停止在 IR 阶段，则继续生成汇编 (后端)
@@ -166,6 +183,8 @@ int main(int argc, char **argv) {
       ofstream fout(argOutputFilename);
       if (not fout.is_open()) {
         cerr << "Failed to open output file: " << argOutputFilename << endl;
+        moduleIR->cleanup(); // 清理模块
+        sysy::cleanupIRPools(); // 清理内存池
         return EXIT_FAILURE;
       }
       fout << asmCode << endl;
@@ -173,6 +192,8 @@ int main(int argc, char **argv) {
     } else {
       cout << asmCode << endl;
     }
+    moduleIR->cleanup(); // 清理模块
+    sysy::cleanupIRPools(); // 清理内存池
     return EXIT_SUCCESS;
   }
 
@@ -181,5 +202,7 @@ int main(int argc, char **argv) {
   cout << "Compilation completed. No output specified (neither -s nor -S). Exiting.\n";
   // return EXIT_SUCCESS; // 或者这里调用一个链接器生成可执行文件
 
+  moduleIR->cleanup(); // 清理模块
+  sysy::cleanupIRPools(); // 清理内存池
   return EXIT_SUCCESS; 
 }