Cache GPU RHS symbols and zero vacuum sources once

.gitignore

@@ -1,6 +1,6 @@
 __pycache__
 GW150914
-GW150914*
+GW150914-origin
 docs
 *.tmp
-.codex
+

AMSS_NCKU_Program.py

@@ -174,14 +174,11 @@ import generate_macrodef
 generate_macrodef.generate_macrodef_h()
 print( " AMSS-NCKU macro file macrodef.h has been generated. " )
      
-generate_macrodef.generate_macrodef_fh()
-print( " AMSS-NCKU macro file macrodef.fh has been generated. " )
-
-generate_macrodef.generate_build_config()
-print( " AMSS-NCKU build config AMSS_NCKU_build.mk has been generated. " )
-
-
-##################################################################
+generate_macrodef.generate_macrodef_fh()
+print( " AMSS-NCKU macro file macrodef.fh has been generated. " )
+
+
+##################################################################
 
 # Compile the AMSS-NCKU program according to user requirements
 
@@ -220,13 +217,11 @@ shutil.copytree(AMSS_NCKU_source_path, AMSS_NCKU_source_copy)
 
 # Copy the generated macro files into the AMSS_NCKU source folder
 
-macrodef_h_path  = os.path.join(File_directory, "macrodef.h")
-macrodef_fh_path = os.path.join(File_directory, "macrodef.fh")
-build_config_path = os.path.join(File_directory, "AMSS_NCKU_build.mk")
-
-shutil.copy2(macrodef_h_path,  AMSS_NCKU_source_copy)
-shutil.copy2(macrodef_fh_path, AMSS_NCKU_source_copy)
-shutil.copy2(build_config_path, AMSS_NCKU_source_copy)
+macrodef_h_path  = os.path.join(File_directory, "macrodef.h") 
+macrodef_fh_path = os.path.join(File_directory, "macrodef.fh") 
+
+shutil.copy2(macrodef_h_path,  AMSS_NCKU_source_copy)
+shutil.copy2(macrodef_fh_path, AMSS_NCKU_source_copy)
 
 # Notes on copying files:
 # shutil.copy2 preserves file metadata such as modification time.

AMSS_NCKU_Program_Plot.py

@@ -1,100 +0,0 @@
-##################################################################
-##
-## AMSS-NCKU Plot-Only Restart Script
-## Author: Xiaoqu / Claude
-## 2026/05/12
-##
-## This script checks for existing output data from AMSS_NCKU_Program.py.
-## If data exists, it skips all computation and goes directly to plotting,
-## saving time when plotting was interrupted.
-## If no data is found, it exits with a message.
-##
-##################################################################
-
-## Guard against re-execution by multiprocessing child processes.
-if __name__ != '__main__':
-    import sys as _sys
-    _sys.exit(0)
-
-
-import os
-import sys
-
-import AMSS_NCKU_Input as input_data
-
-##################################################################
-
-## Construct paths from input configuration
-File_directory         = os.path.join(input_data.File_directory)
-output_directory       = os.path.join(File_directory, "AMSS_NCKU_output")
-binary_results_directory = os.path.join(output_directory, input_data.Output_directory)
-figure_directory       = os.path.join(File_directory, "figure")
-
-##################################################################
-
-## Check whether the required output data files exist
-
-required_files = [
-    os.path.join(binary_results_directory, "bssn_BH.dat"),
-    os.path.join(binary_results_directory, "bssn_ADMQs.dat"),
-    os.path.join(binary_results_directory, "bssn_psi4.dat"),
-    os.path.join(binary_results_directory, "bssn_constraint.dat"),
-]
-
-missing_files = [f for f in required_files if not os.path.exists(f)]
-
-if missing_files:
-    print(" No existing AMSS_NCKU_Program.py output data found. ")
-    print(" The following required files are missing: ")
-    for f in missing_files:
-        print(f"   {f}")
-    print()
-    print(" Please run AMSS_NCKU_Program.py first to generate the simulation data. ")
-    print(" Exiting. ")
-    sys.exit(1)
-
-print(" Found existing AMSS_NCKU_Program.py output data. "          )
-print(" Skipping all computation and going directly to plotting. " )
-print()
-
-## Ensure the figure directory exists (it should, but be safe)
-os.makedirs(figure_directory, exist_ok=True)
-
-##################################################################
-
-## Plot the AMSS-NCKU program results
-
-import plot_xiaoqu
-import plot_GW_strain_amplitude_xiaoqu
-from parallel_plot_helper import run_plot_tasks_parallel
-
-plot_tasks = []
-
-## Plot black hole trajectory
-plot_tasks.append((plot_xiaoqu.generate_puncture_orbit_plot,   (binary_results_directory, figure_directory)))
-plot_tasks.append((plot_xiaoqu.generate_puncture_orbit_plot3D, (binary_results_directory, figure_directory)))
-
-## Plot black hole separation vs. time
-plot_tasks.append((plot_xiaoqu.generate_puncture_distence_plot, (binary_results_directory, figure_directory)))
-
-## Plot gravitational waveforms (psi4 and strain amplitude)
-for i in range(input_data.Detector_Number):
-    plot_tasks.append((plot_xiaoqu.generate_gravitational_wave_psi4_plot, (binary_results_directory, figure_directory, i)))
-    plot_tasks.append((plot_GW_strain_amplitude_xiaoqu.generate_gravitational_wave_amplitude_plot, (binary_results_directory, figure_directory, i)))
-
-## Plot ADM mass evolution
-for i in range(input_data.Detector_Number):
-    plot_tasks.append((plot_xiaoqu.generate_ADMmass_plot, (binary_results_directory, figure_directory, i)))
-
-## Plot Hamiltonian constraint violation over time
-for i in range(input_data.grid_level):
-    plot_tasks.append((plot_xiaoqu.generate_constraint_check_plot, (binary_results_directory, figure_directory, i)))
-
-run_plot_tasks_parallel(plot_tasks)
-
-## Plot stored binary data (runs serially, not in the parallel pool)
-plot_xiaoqu.generate_binary_data_plot(binary_results_directory, figure_directory)
-
-print()
-print(" Plotting completed successfully. ")
-print()

AMSS_NCKU_Verify_ASC26.py

@@ -2,18 +2,13 @@
 """
 AMSS-NCKU GW150914 Simulation Regression Test Script (Comprehensive Version)
 
-Verification Requirements:
-1. RMS errors < 1% for:
-   - 3D Vector Total RMS
-   - X Component RMS
-   - Y Component RMS
-   - Z Component RMS
-2. ADM constraint violation < 2 (Grid Level 0)
-3. The following figure PDFs must match GW150914-origin exactly after rasterization:
-   - ADM_Constraint_Grid_Level_0.pdf
-   - BH_Trajectory_21_XY.pdf
-   - BH_Trajectory_XY.pdf
-   The script also reports the percentage of differing pixels for each figure.
+Verification Requirements:
+1. RMS errors < 1% for:
+   - 3D Vector Total RMS
+   - X Component RMS
+   - Y Component RMS
+   - Z Component RMS
+2. ADM constraint violation < 2 (Grid Level 0)
 
 RMS Calculation Method:
 - Computes trajectory deviation on the XY plane independently for BH1 and BH2
@@ -25,13 +20,9 @@ Default: output_dir = GW150914/AMSS_NCKU_output
 Reference: GW150914-origin (baseline simulation)
 """
 
-import numpy as np
-import sys
-import os
-import shutil
-import subprocess
-import tempfile
-from PIL import Image
+import numpy as np
+import sys
+import os
 
 # ANSI Color Codes
 class Color:
@@ -58,143 +49,17 @@ def load_bh_trajectory(filepath):
     }
 
 
-def load_constraint_data(filepath):
-    """Load constraint violation data"""
-    data = []
+def load_constraint_data(filepath):
+    """Load constraint violation data"""
+    data = []
     with open(filepath, 'r') as f:
         for line in f:
             if line.startswith('#'):
                 continue
             parts = line.split()
             if len(parts) >= 8:
-                data.append([float(x) for x in parts[:8]])
-    return np.array(data)
-
-
-def resolve_figure_dir(path):
-    """Resolve the sibling figure directory from an output or figure path."""
-    normalized = os.path.normpath(path)
-    if os.path.basename(normalized) == "figure":
-        return normalized
-    return os.path.join(os.path.dirname(normalized), "figure")
-
-
-def render_pdf_to_images(pdf_path, dpi=150):
-    """Render a PDF to RGB images using Ghostscript."""
-    gs_path = shutil.which("gs")
-    if gs_path is None:
-        raise RuntimeError("Ghostscript executable 'gs' was not found in PATH")
-
-    with tempfile.TemporaryDirectory(prefix="amss_verify_pdf_") as temp_dir:
-        output_pattern = os.path.join(temp_dir, "page-%03d.ppm")
-        cmd = [
-            gs_path,
-            "-q",
-            "-dSAFER",
-            "-dBATCH",
-            "-dNOPAUSE",
-            "-sDEVICE=ppmraw",
-            f"-r{dpi}",
-            f"-o{output_pattern}",
-            pdf_path
-        ]
-
-        try:
-            subprocess.run(cmd, check=True, stdout=subprocess.DEVNULL, stderr=subprocess.PIPE, text=True)
-        except subprocess.CalledProcessError as exc:
-            message = exc.stderr.strip() or str(exc)
-            raise RuntimeError(f"Failed to render PDF '{pdf_path}': {message}") from exc
-
-        ppm_files = sorted(
-            os.path.join(temp_dir, filename)
-            for filename in os.listdir(temp_dir)
-            if filename.endswith(".ppm")
-        )
-
-        if not ppm_files:
-            raise RuntimeError(f"No rendered pages were produced for '{pdf_path}'")
-
-        images = []
-        for ppm_file in ppm_files:
-            with Image.open(ppm_file) as img:
-                images.append(np.array(img.convert("RGB"), dtype=np.uint8))
-
-        return images
-
-
-def compare_rendered_pages(ref_img, target_img):
-    """Return (different_pixels, total_pixels) for two rendered RGB pages."""
-    ref_h, ref_w = ref_img.shape[:2]
-    tgt_h, tgt_w = target_img.shape[:2]
-    total_pixels = max(ref_h, tgt_h) * max(ref_w, tgt_w)
-
-    if ref_h == tgt_h and ref_w == tgt_w:
-        different_pixels = int(np.count_nonzero(np.any(ref_img != target_img, axis=2)))
-        return different_pixels, total_pixels
-
-    diff_mask = np.ones((max(ref_h, tgt_h), max(ref_w, tgt_w)), dtype=bool)
-    overlap_h = min(ref_h, tgt_h)
-    overlap_w = min(ref_w, tgt_w)
-    overlap_diff = np.any(ref_img[:overlap_h, :overlap_w] != target_img[:overlap_h, :overlap_w], axis=2)
-    diff_mask[:overlap_h, :overlap_w] = overlap_diff
-    different_pixels = int(np.count_nonzero(diff_mask))
-    return different_pixels, total_pixels
-
-
-def compare_pdf_images(ref_pdf, target_pdf, dpi=150, threshold_percent=0.001):
-    """Compare two PDFs by rasterizing them and counting differing pixels."""
-    ref_pages = render_pdf_to_images(ref_pdf, dpi=dpi)
-    target_pages = render_pdf_to_images(target_pdf, dpi=dpi)
-
-    total_pixels = 0
-    different_pixels = 0
-    max_pages = max(len(ref_pages), len(target_pages))
-
-    for page_idx in range(max_pages):
-        if page_idx < len(ref_pages) and page_idx < len(target_pages):
-            page_diff, page_total = compare_rendered_pages(ref_pages[page_idx], target_pages[page_idx])
-        else:
-            existing_page = ref_pages[page_idx] if page_idx < len(ref_pages) else target_pages[page_idx]
-            page_total = existing_page.shape[0] * existing_page.shape[1]
-            page_diff = page_total
-
-        total_pixels += page_total
-        different_pixels += page_diff
-
-    diff_percent = (different_pixels / total_pixels * 100.0) if total_pixels else 0.0
-    return {
-        "different_pixels": different_pixels,
-        "total_pixels": total_pixels,
-        "diff_percent": diff_percent,
-        "pages_ref": len(ref_pages),
-        "pages_target": len(target_pages),
-        "passed": diff_percent < threshold_percent
-    }
-
-
-def compare_required_figures(reference_figure_dir, target_figure_dir):
-    """Compare the required GW150914 figure PDFs."""
-    figure_names = [
-        "ADM_Constraint_Grid_Level_0.pdf",
-        "BH_Trajectory_21_XY.pdf",
-        "BH_Trajectory_XY.pdf"
-    ]
-
-    results = []
-    for figure_name in figure_names:
-        ref_pdf = os.path.join(reference_figure_dir, figure_name)
-        target_pdf = os.path.join(target_figure_dir, figure_name)
-
-        if not os.path.exists(ref_pdf):
-            raise FileNotFoundError(f"Reference figure not found: {ref_pdf}")
-        if not os.path.exists(target_pdf):
-            raise FileNotFoundError(f"Target figure not found: {target_pdf}")
-
-        comparison = compare_pdf_images(ref_pdf, target_pdf)
-        comparison["name"] = figure_name
-        results.append(comparison)
-
-    return results
+                data.append([float(x) for x in parts[:8]])
+    return np.array(data)
 
 def calculate_all_rms_errors(bh_data_ref, bh_data_target):
     """
@@ -300,7 +165,7 @@ def print_rms_results(rms_dict, error, threshold=1.0):
 
     return all_passed
 
-def print_constraint_results(results, threshold=2.0):
+def print_constraint_results(results, threshold=2.0):
     print(f"\n{Color.BOLD}2. ADM Constraint Violation Analysis (Grid Level 0){Color.RESET}")
     print("-" * 65)
 
@@ -315,49 +180,22 @@ def print_constraint_results(results, threshold=2.0):
     print(f"\n   Maximum violation:  {results['max_violation']:.6f}")
     print(f"   Requirement:        < {threshold}")
     print(f"   Status:             {get_status_text(passed)}")
-
-    return passed
-
-
-def print_figure_results(results, threshold_percent=0.001):
-    print(f"\n{Color.BOLD}3. Figure Pixel Comparison (PDF Rasterization){Color.RESET}")
-    print("-" * 65)
-    print(f"   Requirement: < {threshold_percent:.3f}% differing pixels\n")
-
-    all_passed = True
-    for result in results:
-        passed = result["passed"]
-        all_passed = all_passed and passed
-        status = get_status_text(passed)
-        print(f"   {result['name']:32}: {result['diff_percent']:10.6f}%   |   Status: {status}")
-
-        if result["pages_ref"] != result["pages_target"]:
-            print(f"   {'':32}  pages(ref/target): {result['pages_ref']}/{result['pages_target']}")
-
-    return all_passed
-
-
-def print_figure_error(error_message):
-    print(f"\n{Color.BOLD}3. Figure Pixel Comparison (PDF Rasterization){Color.RESET}")
-    print("-" * 65)
-    print(f"   {Color.RED}Error: {error_message}{Color.RESET}")
-    return False
-
-
-def print_summary(rms_passed, constraint_passed, figure_passed):
-    print("\n" + Color.BLUE + Color.BOLD + "=" * 65 + Color.RESET)
-    print(Color.BOLD + "Verification Summary" + Color.RESET)
-    print(Color.BLUE + Color.BOLD + "=" * 65 + Color.RESET)
-
-    all_passed = rms_passed and constraint_passed and figure_passed
-    
-    res_rms = get_status_text(rms_passed)
-    res_con = get_status_text(constraint_passed)
-    res_fig = get_status_text(figure_passed)
-
-    print(f"   [1] Comprehensive RMS check:      {res_rms}")
-    print(f"   [2] ADM constraint check:         {res_con}")
-    print(f"   [3] Figure pixel comparison:      {res_fig}")
+
+    return passed
+
+
+def print_summary(rms_passed, constraint_passed):
+    print("\n" + Color.BLUE + Color.BOLD + "=" * 65 + Color.RESET)
+    print(Color.BOLD + "Verification Summary" + Color.RESET)
+    print(Color.BLUE + Color.BOLD + "=" * 65 + Color.RESET)
+
+    all_passed = rms_passed and constraint_passed
+    
+    res_rms = get_status_text(rms_passed)
+    res_con = get_status_text(constraint_passed)
+
+    print(f"   [1] Comprehensive RMS check:      {res_rms}")
+    print(f"   [2] ADM constraint check:         {res_con}")
     
     final_status = f"{Color.GREEN}{Color.BOLD}ALL CHECKS PASSED{Color.RESET}" if all_passed else f"{Color.RED}{Color.BOLD}SOME CHECKS FAILED{Color.RESET}"
     print(f"\n   Overall result: {final_status}")
@@ -372,14 +210,12 @@ def main():
         script_dir = os.path.dirname(os.path.abspath(__file__))
         target_dir = os.path.join(script_dir, "GW150914/AMSS_NCKU_output")
 
-    script_dir = os.path.dirname(os.path.abspath(__file__))
-    reference_dir = os.path.join(script_dir, "GW150914-origin/AMSS_NCKU_output")
-    target_figure_dir = resolve_figure_dir(target_dir)
-    reference_figure_dir = os.path.join(script_dir, "GW150914-origin/figure")
-
-    bh_file_ref = os.path.join(reference_dir, "bssn_BH.dat")
-    bh_file_target = os.path.join(target_dir, "bssn_BH.dat")
-    constraint_file = os.path.join(target_dir, "bssn_constraint.dat")
+    script_dir = os.path.dirname(os.path.abspath(__file__))
+    reference_dir = os.path.join(script_dir, "GW150914-origin/AMSS_NCKU_output")
+
+    bh_file_ref = os.path.join(reference_dir, "bssn_BH.dat")
+    bh_file_target = os.path.join(target_dir, "bssn_BH.dat")
+    constraint_file = os.path.join(target_dir, "bssn_constraint.dat")
 
     if not os.path.exists(bh_file_ref):
         print(f"{Color.RED}{Color.BOLD}Error:{Color.RESET} Baseline trajectory file not found: {bh_file_ref}")
@@ -391,11 +227,9 @@ def main():
         print(f"{Color.RED}{Color.BOLD}Error:{Color.RESET} Constraint data file not found: {constraint_file}")
         sys.exit(1)
 
-    print_header()
-    print(f"\n{Color.BOLD}Reference (Baseline):{Color.RESET} {Color.BLUE}{reference_dir}{Color.RESET}")
-    print(f"{Color.BOLD}Target (Optimized):  {Color.RESET} {Color.BLUE}{target_dir}{Color.RESET}")
-    print(f"{Color.BOLD}Reference Figures:   {Color.RESET} {Color.BLUE}{reference_figure_dir}{Color.RESET}")
-    print(f"{Color.BOLD}Target Figures:      {Color.RESET} {Color.BLUE}{target_figure_dir}{Color.RESET}")
+    print_header()
+    print(f"\n{Color.BOLD}Reference (Baseline):{Color.RESET} {Color.BLUE}{reference_dir}{Color.RESET}")
+    print(f"{Color.BOLD}Target (Optimized):  {Color.RESET} {Color.BLUE}{target_dir}{Color.RESET}")
 
     bh_data_ref = load_bh_trajectory(bh_file_ref)
     bh_data_target = load_bh_trajectory(bh_file_target)
@@ -405,18 +239,12 @@ def main():
     rms_dict, error = calculate_all_rms_errors(bh_data_ref, bh_data_target)
     rms_passed = print_rms_results(rms_dict, error)
 
-    # Output constraint results
-    constraint_results = analyze_constraint_violation(constraint_data)
-    constraint_passed = print_constraint_results(constraint_results)
-
-    try:
-        figure_results = compare_required_figures(reference_figure_dir, target_figure_dir)
-        figure_passed = print_figure_results(figure_results)
-    except (FileNotFoundError, RuntimeError) as exc:
-        figure_passed = print_figure_error(str(exc))
-
-    all_passed = print_summary(rms_passed, constraint_passed, figure_passed)
-    sys.exit(0 if all_passed else 1)
+    # Output constraint results
+    constraint_results = analyze_constraint_violation(constraint_data)
+    constraint_passed = print_constraint_results(constraint_results)
+
+    all_passed = print_summary(rms_passed, constraint_passed)
+    sys.exit(0 if all_passed else 1)
 
 if __name__ == "__main__":
     main()

AMSS_NCKU_source/ABE.C

@@ -23,22 +23,20 @@ using namespace std;
 #include <mpi.h>
 
 #include "misc.h"
-#include "macrodef.h"
+#include "macrodef.h"
+#ifdef USE_GPU
+extern void bssn_cuda_dump_stage_profile();
+#endif
 
 #ifndef ABEtype
 #error "not define ABEtype"
 #endif
 
-#if (ABEtype == 0)
-
-#ifdef USE_GPU
-#include "bssn_gpu_class.h"
-#else
-#include "bssn_class.h"
-#endif
-
-#elif (ABEtype == 1)
-#include "bssnEScalar_class.h"
+#if (ABEtype == 0)
+#include "bssn_class.h"
+
+#elif (ABEtype == 1)
+#include "bssnEScalar_class.h"
 
 #elif (ABEtype == 2)
 #include "Z4c_class.h"
@@ -474,10 +472,13 @@ int main(int argc, char *argv[])
             cout << endl;
       }
 
-      ADM->Evolve(Steps);
-
-      if (myrank == 0)
-      {
+	      ADM->Evolve(Steps);
+#ifdef USE_GPU
+	      bssn_cuda_dump_stage_profile();
+#endif
+	
+	      if (myrank == 0)
+	      {
             cout << endl;
             cout << " Total Evolve Time: "  << MPI_Wtime() - End_clock   << " seconds!" << endl;
             cout << " Total Running Time: " << MPI_Wtime() - Begin_clock << " seconds!" << endl;

AMSS_NCKU_source/Block.C

@@ -9,8 +9,12 @@
 #include <new>
 using namespace std;
 
-#include "Block.h"
-#include "misc.h"
+#include "Block.h"
+#include "misc.h"
+#ifdef USE_GPU
+#include "bssn_gpu.h"
+#include "bssn_cuda_ops.h"
+#endif
 
 Block::Block(int DIM, int *shapei, double *bboxi, int ranki, int ingfsi, int fngfsi, int levi, const int cgpui) : rank(ranki), ingfs(ingfsi), fngfs(fngfsi), lev(levi), cgpu(cgpui)
 {
@@ -95,14 +99,19 @@ Block::Block(int DIM, int *shapei, double *bboxi, int ranki, int ingfsi, int fng
   }
 #endif
 }
-Block::~Block()
-{
-  int myrank;
-  MPI_Comm_rank(MPI_COMM_WORLD, &myrank);
-  if (myrank == rank)
-  {
-    for (int i = 0; i < dim; i++)
-      delete[] X[i];
+Block::~Block()
+{
+  int myrank;
+  MPI_Comm_rank(MPI_COMM_WORLD, &myrank);
+  if (myrank == rank)
+  {
+#ifdef USE_GPU
+    bssn_gpu_clear_cached_device_buffers();
+    bssn_cuda_release_rk4_caches();
+    bssn_cuda_release_interp_caches();
+#endif
+    for (int i = 0; i < dim; i++)
+      delete[] X[i];
     for (int i = 0; i < ingfs; i++)
       free(igfs[i]);
     delete[] igfs;

AMSS_NCKU_source/MPatch.C

@@ -2,29 +2,100 @@
 #include <iostream>
 #include <iomanip>
 #include <fstream>
-#include <cstdlib>
-#include <cstdio>
-#include <string>
-#include <cmath>
-#include <new>
-#include <vector>
-using namespace std;
+#include <cstdlib>
+#include <cstdio>
+#include <string>
+#include <cmath>
+#include <new>
+#include <map>
+#include <vector>
+using namespace std;
 
-#include "misc.h"
-#include "MPatch.h"
-#include "Parallel.h"
-#include "fmisc.h"
-#ifdef INTERP_LB_PROFILE
-#include "interp_lb_profile.h"
-#endif
-
-namespace
-{
-struct InterpBlockView
-{
-  Block *bp;
-  double llb[dim];
-  double uub[dim];
+#include "misc.h"
+#include "MPatch.h"
+#include "Parallel.h"
+#include "fmisc.h"
+#include "bssn_cuda_ops.h"
+#ifdef INTERP_LB_PROFILE
+#include "interp_lb_profile.h"
+#endif
+
+#if defined(__GNUC__) || defined(__clang__)
+extern int bssn_cuda_interp_points_batch(const int *ex,
+                                         const double *X, const double *Y, const double *Z,
+                                         const double *const *fields,
+                                         const double *soa_flat,
+                                         int num_var,
+                                         const double *px, const double *py, const double *pz,
+                                         int num_points,
+                                         int ordn,
+                                         int symmetry,
+                                         double *out) __attribute__((weak));
+#endif
+
+namespace
+{
+struct InterpVarDesc
+{
+  int sgfn;
+  double soa[dim];
+};
+
+struct InterpPlanKey
+{
+  const Patch *patch;
+  const double *x;
+  const double *y;
+  const double *z;
+  int NN;
+  int Symmetry;
+  int myrank;
+};
+
+struct InterpPlanKeyLess
+{
+  bool operator()(const InterpPlanKey &lhs, const InterpPlanKey &rhs) const
+  {
+    if (lhs.patch != rhs.patch) return lhs.patch < rhs.patch;
+    if (lhs.x != rhs.x) return lhs.x < rhs.x;
+    if (lhs.y != rhs.y) return lhs.y < rhs.y;
+    if (lhs.z != rhs.z) return lhs.z < rhs.z;
+    if (lhs.NN != rhs.NN) return lhs.NN < rhs.NN;
+    if (lhs.Symmetry != rhs.Symmetry) return lhs.Symmetry < rhs.Symmetry;
+    return lhs.myrank < rhs.myrank;
+  }
+};
+
+struct CachedInterpPlan
+{
+  int nblocks;
+  vector<int> owner_rank;
+  vector<int> owner_block;
+  vector<vector<int> > block_points;
+  vector<vector<double> > block_px;
+  vector<vector<double> > block_py;
+  vector<vector<double> > block_pz;
+
+  CachedInterpPlan() : nblocks(0) {}
+};
+
+struct CachedInterpPlanEntry
+{
+  bool valid;
+  InterpPlanKey key;
+  vector<double> xvals;
+  vector<double> yvals;
+  vector<double> zvals;
+  CachedInterpPlan plan;
+
+  CachedInterpPlanEntry() : valid(false) {}
+};
+
+struct InterpBlockView
+{
+  Block *bp;
+  double llb[dim];
+  double uub[dim];
 };
 
 struct BlockBinIndex
@@ -154,10 +225,10 @@ void build_block_bin_index(Patch *patch, const double *DH, BlockBinIndex &index)
   index.valid = true;
 }
 
-int find_block_index_for_point(const BlockBinIndex &index, const double *pox, const double *DH)
-{
-  if (!index.valid)
-    return -1;
+int find_block_index_for_point(const BlockBinIndex &index, const double *pox, const double *DH)
+{
+  if (!index.valid)
+    return -1;
 
   const int bx = coord_to_bin(pox[0], index.lo[0], index.inv[0], index.bins[0]);
   const int by = coord_to_bin(pox[1], index.lo[1], index.inv[1], index.bins[1]);
@@ -175,13 +246,314 @@ int find_block_index_for_point(const BlockBinIndex &index, const double *pox, co
   for (size_t bi = 0; bi < index.views.size(); bi++)
     if (point_in_block_view(index.views[bi], pox, DH))
       return int(bi);
-
-  return -1;
-}
-} // namespace
-
-Patch::Patch(int DIM, int *shapei, double *bboxi, int levi, bool buflog, int Symmetry) : lev(levi)
-{
+
+  return -1;
+}
+
+void collect_interp_vars(MyList<var> *VarList, vector<InterpVarDesc> &vars)
+{
+  vars.clear();
+  MyList<var> *varl = VarList;
+  while (varl)
+  {
+    InterpVarDesc desc;
+    desc.sgfn = varl->data->sgfn;
+    for (int d = 0; d < dim; ++d)
+      desc.soa[d] = varl->data->SoA[d];
+    vars.push_back(desc);
+    varl = varl->next;
+  }
+}
+
+bool should_try_cuda_interp(int ordn, int num_points, int num_var)
+{
+#if defined(__GNUC__) || defined(__clang__)
+  if (!bssn_cuda_interp_points_batch)
+    return false;
+#else
+  return false;
+#endif
+  if (ordn != 6)
+    return false;
+  if (num_points < 32)
+    return false;
+  return num_points * num_var >= 256;
+}
+
+CachedInterpPlanEntry &interp_plan_cache_entry()
+{
+  static CachedInterpPlanEntry cache;
+  return cache;
+}
+
+bool same_interp_plan_key(const InterpPlanKey &lhs, const InterpPlanKey &rhs)
+{
+  return lhs.patch == rhs.patch &&
+         lhs.NN == rhs.NN &&
+         lhs.Symmetry == rhs.Symmetry &&
+         lhs.myrank == rhs.myrank;
+}
+
+bool same_interp_plan_points(const CachedInterpPlanEntry &cache, int NN, double **XX)
+{
+  if (static_cast<int>(cache.xvals.size()) != NN ||
+      static_cast<int>(cache.yvals.size()) != NN ||
+      static_cast<int>(cache.zvals.size()) != NN)
+    return false;
+
+  for (int j = 0; j < NN; ++j)
+  {
+    if (cache.xvals[j] != XX[0][j] ||
+        cache.yvals[j] != XX[1][j] ||
+        cache.zvals[j] != XX[2][j])
+      return false;
+  }
+  return true;
+}
+
+CachedInterpPlan &get_cached_interp_plan(Patch *patch,
+                                         int NN, double **XX,
+                                         int Symmetry, int myrank,
+                                         const double *DH,
+                                         const BlockBinIndex &block_index,
+                                         bool report_bounds_here,
+                                         bool allow_missing_points)
+{
+  InterpPlanKey key;
+  key.patch = patch;
+  key.x = XX[0];
+  key.y = XX[1];
+  key.z = XX[2];
+  key.NN = NN;
+  key.Symmetry = Symmetry;
+  key.myrank = myrank;
+
+  CachedInterpPlanEntry &cache = interp_plan_cache_entry();
+  if (cache.valid &&
+      same_interp_plan_key(cache.key, key) &&
+      same_interp_plan_points(cache, NN, XX) &&
+      cache.plan.nblocks == static_cast<int>(block_index.views.size()))
+    return cache.plan;
+
+  cache.valid = true;
+  cache.key = key;
+  cache.xvals.assign(XX[0], XX[0] + NN);
+  cache.yvals.assign(XX[1], XX[1] + NN);
+  cache.zvals.assign(XX[2], XX[2] + NN);
+  cache.plan = CachedInterpPlan();
+  CachedInterpPlan &plan = cache.plan;
+  plan.nblocks = static_cast<int>(block_index.views.size());
+  plan.owner_rank.assign(NN, -1);
+  plan.owner_block.assign(NN, -1);
+  plan.block_points.resize(plan.nblocks);
+  plan.block_px.resize(plan.nblocks);
+  plan.block_py.resize(plan.nblocks);
+  plan.block_pz.resize(plan.nblocks);
+
+  for (int j = 0; j < NN; ++j)
+  {
+    double pox[dim];
+    for (int i = 0; i < dim; ++i)
+    {
+      pox[i] = XX[i][j];
+      if (report_bounds_here &&
+          (XX[i][j] < patch->bbox[i] + patch->lli[i] * DH[i] ||
+           XX[i][j] > patch->bbox[dim + i] - patch->uui[i] * DH[i]))
+      {
+        cout << "Patch::Interp_Points: point (";
+        for (int k = 0; k < dim; ++k)
+        {
+          cout << XX[k][j];
+          if (k < dim - 1)
+            cout << ",";
+          else
+            cout << ") is out of current Patch." << endl;
+        }
+        MPI_Abort(MPI_COMM_WORLD, 1);
+      }
+    }
+
+    const int block_i = find_block_index_for_point(block_index, pox, DH);
+    if (block_i >= 0)
+    {
+      Block *BP = block_index.views[block_i].bp;
+      plan.owner_rank[j] = BP->rank;
+      plan.owner_block[j] = block_i;
+      if (BP->rank == myrank)
+      {
+        plan.block_points[block_i].push_back(j);
+        plan.block_px[block_i].push_back(XX[0][j]);
+        plan.block_py[block_i].push_back(XX[1][j]);
+        plan.block_pz[block_i].push_back(XX[2][j]);
+      }
+    }
+  }
+
+  if (!allow_missing_points && report_bounds_here)
+  {
+    for (int j = 0; j < NN; ++j)
+    {
+      if (plan.owner_rank[j] >= 0)
+        continue;
+      cout << "ERROR: Patch::Interp_Points fails to find point (";
+      for (int d = 0; d < dim; ++d)
+      {
+        cout << XX[d][j];
+        if (d < dim - 1)
+          cout << ",";
+        else
+          cout << ")";
+      }
+      cout << " on Patch (";
+      for (int d = 0; d < dim; ++d)
+      {
+        cout << patch->bbox[d] << "+" << patch->lli[d] * DH[d];
+        if (d < dim - 1)
+          cout << ",";
+        else
+          cout << ")--";
+      }
+      cout << "(";
+      for (int d = 0; d < dim; ++d)
+      {
+        cout << patch->bbox[dim + d] << "-" << patch->uui[d] * DH[d];
+        if (d < dim - 1)
+          cout << ",";
+        else
+          cout << ")" << endl;
+      }
+      MPI_Abort(MPI_COMM_WORLD, 1);
+    }
+  }
+
+  return plan;
+}
+
+void release_interp_plan_cache_internal()
+{
+  CachedInterpPlanEntry &cache = interp_plan_cache_entry();
+  cache.valid = false;
+  cache.xvals.clear();
+  cache.yvals.clear();
+  cache.zvals.clear();
+  cache.plan = CachedInterpPlan();
+}
+
+bool run_cuda_interp_for_block(Block *BP,
+                               const vector<InterpVarDesc> &vars,
+                               const vector<int> &point_ids,
+                               const vector<double> &px,
+                               const vector<double> &py,
+                               const vector<double> &pz,
+                               double *Shellf,
+                               int num_var,
+                               int ordn,
+                               int Symmetry)
+{
+  if (!should_try_cuda_interp(ordn, static_cast<int>(point_ids.size()), num_var))
+    return false;
+
+  vector<const double *> field_ptrs(num_var);
+  vector<double> soa_flat(3 * num_var);
+  for (int v = 0; v < num_var; ++v)
+  {
+    field_ptrs[v] = BP->fgfs[vars[v].sgfn];
+    for (int d = 0; d < dim; ++d)
+      soa_flat[3 * v + d] = vars[v].soa[d];
+  }
+
+  const int npts = static_cast<int>(point_ids.size());
+  vector<double> out(static_cast<size_t>(npts) * static_cast<size_t>(num_var));
+  if (bssn_cuda_interp_points_batch(BP->shape,
+                                    BP->X[0], BP->X[1], BP->X[2],
+                                    field_ptrs.data(),
+                                    soa_flat.data(),
+                                    num_var,
+                                    px.data(), py.data(), pz.data(),
+                                    npts,
+                                    ordn,
+                                    Symmetry,
+                                    out.data()) != 0)
+  {
+    return false;
+  }
+
+  for (int p = 0; p < npts; ++p)
+  {
+    const int j = point_ids[p];
+    memcpy(Shellf + j * num_var, out.data() + p * num_var, sizeof(double) * num_var);
+  }
+  return true;
+}
+
+void run_cpu_interp_for_block(Block *BP,
+                              const vector<InterpVarDesc> &vars,
+                              const vector<int> &point_ids,
+                              const vector<double> &px,
+                              const vector<double> &py,
+                              const vector<double> &pz,
+                              double *Shellf,
+                              int num_var,
+                              int ordn,
+                              int Symmetry)
+{
+  for (size_t p = 0; p < point_ids.size(); ++p)
+  {
+    const int j = point_ids[p];
+    double x = px[p];
+    double y = py[p];
+    double z = pz[p];
+    int ordn_local = ordn;
+    int symmetry_local = Symmetry;
+    for (int v = 0; v < num_var; ++v)
+    {
+      f_global_interp(BP->shape, BP->X[0], BP->X[1], BP->X[2],
+                      BP->fgfs[vars[v].sgfn], Shellf[j * num_var + v],
+                      x, y, z, ordn_local, const_cast<double *>(vars[v].soa), symmetry_local);
+    }
+  }
+}
+
+void interpolate_owned_points(MyList<var> *VarList,
+                              double *Shellf, int Symmetry,
+                              int ordn,
+                              const BlockBinIndex &block_index,
+                              const CachedInterpPlan &plan)
+{
+  vector<InterpVarDesc> vars;
+  collect_interp_vars(VarList, vars);
+  const int num_var = static_cast<int>(vars.size());
+
+  for (size_t bi = 0; bi < plan.block_points.size(); ++bi)
+  {
+    if (plan.block_points[bi].empty())
+      continue;
+
+    Block *BP = block_index.views[bi].bp;
+    bool done = run_cuda_interp_for_block(BP, vars,
+                                          plan.block_points[bi],
+                                          plan.block_px[bi],
+                                          plan.block_py[bi],
+                                          plan.block_pz[bi],
+                                          Shellf, num_var, ordn, Symmetry);
+    if (!done)
+      run_cpu_interp_for_block(BP, vars,
+                               plan.block_points[bi],
+                               plan.block_px[bi],
+                               plan.block_py[bi],
+                               plan.block_pz[bi],
+                               Shellf, num_var, ordn, Symmetry);
+  }
+}
+} // namespace
+
+void patch_release_interp_plan_cache()
+{
+  release_interp_plan_cache_internal();
+}
+
+Patch::Patch(int DIM, int *shapei, double *bboxi, int levi, bool buflog, int Symmetry) : lev(levi)
+{
 
   int hbuffer_width = buffer_width;
   if (lev == 0)
@@ -523,60 +895,15 @@ void Patch::Interp_Points(MyList<var> *VarList,
 
   memset(Shellf, 0, sizeof(double) * NN * num_var);
 
-  // owner_rank[j] records which MPI rank owns point j
-  // All ranks traverse the same block list so they all agree on ownership
-  int *owner_rank;
-  owner_rank = new int[NN];
-  for (int j = 0; j < NN; j++)
-    owner_rank[j] = -1;
-
-  double DH[dim];
-  for (int i = 0; i < dim; i++)
-    DH[i] = getdX(i);
-  BlockBinIndex block_index;
-  build_block_bin_index(this, DH, block_index);
-
-  for (int j = 0; j < NN; j++) // run along points
-  {
-    double pox[dim];
-    for (int i = 0; i < dim; i++)
-    {
-      pox[i] = XX[i][j];
-      if (myrank == 0 && (XX[i][j] < bbox[i] + lli[i] * DH[i] || XX[i][j] > bbox[dim + i] - uui[i] * DH[i]))
-      {
-        cout << "Patch::Interp_Points: point (";
-        for (int k = 0; k < dim; k++)
-        {
-          cout << XX[k][j];
-          if (k < dim - 1)
-            cout << ",";
-          else
-            cout << ") is out of current Patch." << endl;
-        }
-        MPI_Abort(MPI_COMM_WORLD, 1);
-      }
-    }
-
-    const int block_i = find_block_index_for_point(block_index, pox, DH);
-    if (block_i >= 0)
-    {
-      Block *BP = block_index.views[block_i].bp;
-      owner_rank[j] = BP->rank;
-      if (myrank == BP->rank)
-      {
-        //---> interpolation
-        varl = VarList;
-        int k = 0;
-        while (varl) // run along variables
-        {
-          f_global_interp(BP->shape, BP->X[0], BP->X[1], BP->X[2], BP->fgfs[varl->data->sgfn], Shellf[j * num_var + k],
-                          pox[0], pox[1], pox[2], ordn, varl->data->SoA, Symmetry);
-          varl = varl->next;
-          k++;
-        }
-      }
-    }
-  }
+  double DH[dim];
+  for (int i = 0; i < dim; i++)
+    DH[i] = getdX(i);
+  BlockBinIndex block_index;
+  build_block_bin_index(this, DH, block_index);
+  CachedInterpPlan &plan = get_cached_interp_plan(this, NN, XX, Symmetry, myrank, DH, block_index, myrank == 0, false);
+  const int *owner_rank = plan.owner_rank.data();
+
+  interpolate_owned_points(VarList, Shellf, Symmetry, ordn, block_index, plan);
 
   // Replace MPI_Allreduce with per-owner MPI_Bcast:
   // Group consecutive points by owner rank and broadcast each group.
@@ -631,9 +958,8 @@ void Patch::Interp_Points(MyList<var> *VarList,
       MPI_Bcast(Shellf + jstart * num_var, count, MPI_DOUBLE, cur_owner, MPI_COMM_WORLD);
     }
   }
-
-  delete[] owner_rank;
-}
+
+}
 void Patch::Interp_Points(MyList<var> *VarList,
                           int NN, double **XX,
                           double *Shellf, int Symmetry,
@@ -661,102 +987,22 @@ void Patch::Interp_Points(MyList<var> *VarList,
 
   memset(Shellf, 0, sizeof(double) * NN * num_var);
 
-  // owner_rank[j] records which MPI rank owns point j
-  int *owner_rank;
-  owner_rank = new int[NN];
-  for (int j = 0; j < NN; j++)
-    owner_rank[j] = -1;
-
-  double DH[dim];
-  for (int i = 0; i < dim; i++)
-    DH[i] = getdX(i);
-  BlockBinIndex block_index;
-  build_block_bin_index(this, DH, block_index);
-
-  // --- Interpolation phase (identical to original) ---
-  for (int j = 0; j < NN; j++)
-  {
-    double pox[dim];
-    for (int i = 0; i < dim; i++)
-    {
-      pox[i] = XX[i][j];
-      if (myrank == 0 && (XX[i][j] < bbox[i] + lli[i] * DH[i] || XX[i][j] > bbox[dim + i] - uui[i] * DH[i]))
-      {
-        cout << "Patch::Interp_Points: point (";
-        for (int k = 0; k < dim; k++)
-        {
-          cout << XX[k][j];
-          if (k < dim - 1)
-            cout << ",";
-          else
-            cout << ") is out of current Patch." << endl;
-        }
-        MPI_Abort(MPI_COMM_WORLD, 1);
-      }
-    }
-
-    const int block_i = find_block_index_for_point(block_index, pox, DH);
-    if (block_i >= 0)
-    {
-      Block *BP = block_index.views[block_i].bp;
-      owner_rank[j] = BP->rank;
-      if (myrank == BP->rank)
-      {
-        varl = VarList;
-        int k = 0;
-        while (varl)
-        {
-          f_global_interp(BP->shape, BP->X[0], BP->X[1], BP->X[2], BP->fgfs[varl->data->sgfn], Shellf[j * num_var + k],
-                          pox[0], pox[1], pox[2], ordn, varl->data->SoA, Symmetry);
-          varl = varl->next;
-          k++;
-        }
-      }
-    }
-  }
+  double DH[dim];
+  for (int i = 0; i < dim; i++)
+    DH[i] = getdX(i);
+  BlockBinIndex block_index;
+  build_block_bin_index(this, DH, block_index);
+  CachedInterpPlan &plan = get_cached_interp_plan(this, NN, XX, Symmetry, myrank, DH, block_index, myrank == 0, false);
+  const int *owner_rank = plan.owner_rank.data();
+
+  interpolate_owned_points(VarList, Shellf, Symmetry, ordn, block_index, plan);
 
 #ifdef INTERP_LB_PROFILE
   double t_interp_end = MPI_Wtime();
   double t_interp_local = t_interp_end - t_interp_start;
 #endif
 
-  // --- Error check for unfound points ---
-  for (int j = 0; j < NN; j++)
-  {
-    if (owner_rank[j] < 0 && myrank == 0)
-    {
-      cout << "ERROR: Patch::Interp_Points fails to find point (";
-      for (int d = 0; d < dim; d++)
-      {
-        cout << XX[d][j];
-        if (d < dim - 1)
-          cout << ",";
-        else
-          cout << ")";
-      }
-      cout << " on Patch (";
-      for (int d = 0; d < dim; d++)
-      {
-        cout << bbox[d] << "+" << lli[d] * DH[d];
-        if (d < dim - 1)
-          cout << ",";
-        else
-          cout << ")--";
-      }
-      cout << "(";
-      for (int d = 0; d < dim; d++)
-      {
-        cout << bbox[dim + d] << "-" << uui[d] * DH[d];
-        if (d < dim - 1)
-          cout << ",";
-        else
-          cout << ")" << endl;
-      }
-      MPI_Abort(MPI_COMM_WORLD, 1);
-    }
-  }
-
-  // --- Targeted point-to-point communication phase ---
+  // --- Targeted point-to-point communication phase ---
   // Compute consumer_rank[j] using the same deterministic formula as surface_integral
   int *consumer_rank = new int[NN];
   {
@@ -873,9 +1119,8 @@ void Patch::Interp_Points(MyList<var> *VarList,
   delete[] send_offset;
   delete[] recv_offset;
   delete[] send_count;
-  delete[] recv_count;
-  delete[] consumer_rank;
-  delete[] owner_rank;
+  delete[] recv_count;
+  delete[] consumer_rank;
 
 #ifdef INTERP_LB_PROFILE
   {
@@ -923,64 +1168,20 @@ void Patch::Interp_Points(MyList<var> *VarList,
 
   memset(Shellf, 0, sizeof(double) * NN * num_var);
 
-  // owner_rank[j] stores the global rank that owns point j
-  int *owner_rank;
-  owner_rank = new int[NN];
-  for (int j = 0; j < NN; j++)
-    owner_rank[j] = -1;
+  // Build global-to-local rank translation for Comm_here
+  MPI_Group world_group, local_group;
+  MPI_Comm_group(MPI_COMM_WORLD, &world_group);
+  MPI_Comm_group(Comm_here, &local_group);
 
-  // Build global-to-local rank translation for Comm_here
-  MPI_Group world_group, local_group;
-  MPI_Comm_group(MPI_COMM_WORLD, &world_group);
-  MPI_Comm_group(Comm_here, &local_group);
-
-  double DH[dim];
-  for (int i = 0; i < dim; i++)
-    DH[i] = getdX(i);
-  BlockBinIndex block_index;
-  build_block_bin_index(this, DH, block_index);
-
-  for (int j = 0; j < NN; j++) // run along points
-  {
-    double pox[dim];
-    for (int i = 0; i < dim; i++)
-    {
-      pox[i] = XX[i][j];
-      if (lmyrank == 0 && (XX[i][j] < bbox[i] + lli[i] * DH[i] || XX[i][j] > bbox[dim + i] - uui[i] * DH[i]))
-      {
-        cout << "Patch::Interp_Points: point (";
-        for (int k = 0; k < dim; k++)
-        {
-          cout << XX[k][j];
-          if (k < dim - 1)
-            cout << ",";
-          else
-            cout << ") is out of current Patch." << endl;
-        }
-        MPI_Abort(MPI_COMM_WORLD, 1);
-      }
-    }
-
-    const int block_i = find_block_index_for_point(block_index, pox, DH);
-    if (block_i >= 0)
-    {
-      Block *BP = block_index.views[block_i].bp;
-      owner_rank[j] = BP->rank;
-      if (myrank == BP->rank)
-      {
-        //---> interpolation
-        varl = VarList;
-        int k = 0;
-        while (varl) // run along variables
-        {
-          f_global_interp(BP->shape, BP->X[0], BP->X[1], BP->X[2], BP->fgfs[varl->data->sgfn], Shellf[j * num_var + k],
-                          pox[0], pox[1], pox[2], ordn, varl->data->SoA, Symmetry);
-          varl = varl->next;
-          k++;
-        }
-      }
-    }
-  }
+  double DH[dim];
+  for (int i = 0; i < dim; i++)
+    DH[i] = getdX(i);
+  BlockBinIndex block_index;
+  build_block_bin_index(this, DH, block_index);
+  CachedInterpPlan &plan = get_cached_interp_plan(this, NN, XX, Symmetry, myrank, DH, block_index, lmyrank == 0, true);
+  const int *owner_rank = plan.owner_rank.data();
+
+  interpolate_owned_points(VarList, Shellf, Symmetry, ordn, block_index, plan);
 
   // Collect unique global owner ranks and translate to local ranks in Comm_here
   // Then broadcast each owner's points via MPI_Bcast on Comm_here
@@ -1008,10 +1209,9 @@ void Patch::Interp_Points(MyList<var> *VarList,
     }
   }
 
-  MPI_Group_free(&world_group);
-  MPI_Group_free(&local_group);
-  delete[] owner_rank;
-}
+  MPI_Group_free(&world_group);
+  MPI_Group_free(&local_group);
+}
 void Patch::checkBlock()
 {
   int myrank;

AMSS_NCKU_source/MPatch.h

@@ -8,7 +8,7 @@
 #include "var.h"
 #include "macrodef.h" //need dim here; Vertex or Cell; ghost_width
 
-class Patch
+class Patch
 {
 
 public:
@@ -50,6 +50,8 @@ public:
                          double *Shellf, int Symmetry, MPI_Comm Comm_here);
    void Find_Maximum(MyList<var> *VarList, double *XX,
                      double *Shellf, MPI_Comm Comm_here);
-};
-
-#endif /* PATCH_H */
+};
+
+void patch_release_interp_plan_cache();
+
+#endif /* PATCH_H */

AMSS_NCKU_source/Parallel.h

@@ -89,9 +89,12 @@ namespace Parallel
   void transfermix(MyList<gridseg> **src, MyList<gridseg> **dst,
                    MyList<var> *VarList1 /* source */, MyList<var> *VarList2 /*target */,
                    int Symmetry);
-  void Sync(Patch *Pat, MyList<var> *VarList, int Symmetry);
-  void Sync(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry);
-  void Sync_merged(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry);
+  void Sync(Patch *Pat, MyList<var> *VarList, int Symmetry);
+  void Sync(Patch *Pat, MyList<var> *VarList, int Symmetry, const char *context);
+  void Sync(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry);
+  void Sync(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry, const char *context);
+  void Sync_merged(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry);
+  void Sync_merged(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry, const char *context);
 
   struct SyncCache {
     bool valid;
@@ -105,12 +108,13 @@ namespace Parallel
     int *send_buf_caps;
     int *recv_buf_caps;
     MPI_Request *reqs;
-    MPI_Status *stats;
-    int max_reqs;
-    bool lengths_valid;
-    int *tc_req_node;
-    int *tc_req_is_recv;
-    int *tc_completed;
+    MPI_Status *stats;
+    int max_reqs;
+    bool lengths_valid;
+    int lengths_var_count;
+    int *tc_req_node;
+    int *tc_req_is_recv;
+    int *tc_completed;
     SyncCache();
     void invalidate();
     void destroy();
@@ -121,19 +125,20 @@ namespace Parallel
                        MyList<var> *VarList1, MyList<var> *VarList2,
                        int Symmetry, SyncCache &cache);
 
-  struct AsyncSyncState {
-    int req_no;
-    bool active;
-    int *req_node;
-    int *req_is_recv;
-    int pending_recv;
-    AsyncSyncState() : req_no(0), active(false), req_node(0), req_is_recv(0), pending_recv(0) {}
-  };
+  struct AsyncSyncState {
+    int req_no;
+    bool active;
+    int mpi_tag;
+    int *req_node;
+    int *req_is_recv;
+    int pending_recv;
+    AsyncSyncState() : req_no(0), active(false), mpi_tag(0), req_node(0), req_is_recv(0), pending_recv(0) {}
+  };
 
-  void Sync_start(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry,
-                  SyncCache &cache, AsyncSyncState &state);
-  void Sync_finish(SyncCache &cache, AsyncSyncState &state,
-                   MyList<var> *VarList, int Symmetry);
+  void Sync_start(MyList<Patch> *PatL, MyList<var> *VarList, int Symmetry,
+                  SyncCache &cache, AsyncSyncState &state);
+  void Sync_finish(SyncCache &cache, AsyncSyncState &state,
+                   MyList<var> *VarList, int Symmetry, bool unpack_to_host = true);
   void OutBdLow2Hi(Patch *Patc, Patch *Patf,
                    MyList<var> *VarList1 /* source */, MyList<var> *VarList2 /* target */,
                    int Symmetry);
@@ -179,13 +184,12 @@ namespace Parallel
   MyList<Parallel::gridseg> *clone_gsl(MyList<Parallel::gridseg> *p, bool first_only);
   MyList<Parallel::gridseg> *build_bulk_gsl(Patch *Pat); // similar to build_owned_gsl0 but does not care rank issue
   MyList<Parallel::gridseg> *build_bulk_gsl(Block *bp, Patch *Pat);
-  void build_PhysBD_gstl(Patch *Pat, MyList<Parallel::gridseg> *srci, MyList<Parallel::gridseg> *dsti,
-                         MyList<Parallel::gridseg> **out_src, MyList<Parallel::gridseg> **out_dst);
-  void PeriodicBD(Patch *Pat, MyList<var> *VarList, int Symmetry);
-  double L2Norm(Patch *Pat, var *vf);
-  void L2Norm7(Patch *Pat, var **vf, double *norms);
-  void checkgsl(MyList<Parallel::gridseg> *pp, bool first_only);
-  void checkvarl(MyList<var> *pp, bool first_only);
+  void build_PhysBD_gstl(Patch *Pat, MyList<Parallel::gridseg> *srci, MyList<Parallel::gridseg> *dsti,
+                         MyList<Parallel::gridseg> **out_src, MyList<Parallel::gridseg> **out_dst);
+  void PeriodicBD(Patch *Pat, MyList<var> *VarList, int Symmetry);
+  double L2Norm(Patch *Pat, var *vf);
+  void checkgsl(MyList<Parallel::gridseg> *pp, bool first_only);
+  void checkvarl(MyList<var> *pp, bool first_only);
   MyList<Parallel::gridseg> *divide_gsl(MyList<Parallel::gridseg> *p, Patch *Pat);
   MyList<Parallel::gridseg> *divide_gs(MyList<Parallel::gridseg> *p, Patch *Pat);
   void prepare_inter_time_level(Patch *Pat,
@@ -217,12 +221,11 @@ namespace Parallel
   void aligncheck(double *bbox0, double *bboxl, int lev, double *DH0, int *shape);
   bool point_locat_gsl(double *pox, MyList<Parallel::gridseg> *gsl);
   void checkpatchlist(MyList<Patch> *PatL, bool buflog);
-
-  double L2Norm(Patch *Pat, var *vf, MPI_Comm Comm_here);
-  void L2Norm7(Patch *Pat, var **vf, double *norms, MPI_Comm Comm_here);
-  bool PatList_Interp_Points(MyList<Patch> *PatL, MyList<var> *VarList,
-                             int NN, double **XX,
-                             double *Shellf, int Symmetry, MPI_Comm Comm_here);
+
+  double L2Norm(Patch *Pat, var *vf, MPI_Comm Comm_here);
+  bool PatList_Interp_Points(MyList<Patch> *PatL, MyList<var> *VarList,
+                             int NN, double **XX,
+                             double *Shellf, int Symmetry, MPI_Comm Comm_here);
 #if (PSTR == 1 || PSTR == 2 || PSTR == 3)
   MyList<Block> *distribute(MyList<Patch> *PatchLIST, int cpusize, int ingfsi, int fngfsi,
                             bool periodic, int start_rank, int end_rank, int nodes = 0);

AMSS_NCKU_source/ShellPatch.h

@@ -102,16 +102,6 @@ public:
                  //-1: means no dumy dimension at all; 0: means rho; 1: means sigma
    };
 
-   // Thread-safe search result (no pointers to shared mutable state)
-   struct PointSearchResult
-   {
-      bool found;
-      Block *Bg;
-      double gx, gy, gz; // global Cartesian coordinates
-      double lx, ly, lz; // local coordinates within the found block
-      int ssst;          // source shell-patch type (-1 = Cartesian)
-   };
-
    int myrank;
    int shape[dim];   // for (rho, sigma, R), for rho and sigma means number of points for every pi/2
    double Rrange[2]; // for Rmin and Rmax
@@ -185,12 +175,6 @@ public:
                          MyList<Patch> *Pp, double CDH[dim], MyList<pointstru> *pss);
    bool prolongpointstru(MyList<pointstru> *&psul, bool ssyn, int tsst, MyList<ss_patch> *sPp, double DH[dim],
                          MyList<Patch> *Pp, double CDH[dim], double x, double y, double z, int Symmetry, int rank_in);
-   // Read-only point search — thread-safe (no shared mutable state modified)
-   PointSearchResult prolongpointstru_search(bool ssyn, int tsst, MyList<ss_patch> *sPp, double DH[dim],
-                                             MyList<Patch> *Pp, double CDH[dim], double x, double y, double z,
-                                             int Symmetry, int rank_in);
-   // Append a search result to a linked list — use inside omp critical section
-   void prolongpointstru_append(MyList<pointstru> *&psul, const PointSearchResult &sr, int tsst);
    void setupintintstuff(int cpusize, MyList<Patch> *CPatL, int Symmetry);
    void intertransfer(MyList<pointstru> **src, MyList<pointstru> **dst,
                       MyList<var> *VarList1 /* source */, MyList<var> *VarList2 /*target */,
@@ -214,7 +198,6 @@ public:
    void write_Pablo_file_ss(int *ext, double xmin, double xmax, double ymin, double ymax, double zmin, double zmax,
                             char *filename, int sst);
    double L2Norm(var *vf);
-   void L2Norm7(var **vf, double *norms);
    void Find_Maximum(MyList<var> *VarList, double *XX, double *Shellf);
 };
 

AMSS_NCKU_source/Z4c_rhs.f90

@@ -94,31 +94,29 @@
                Hcon,Mxcon,Mycon,Mzcon,Gmxcon,Gmycon,Gmzcon,                    &            
                Symmetry,Lev,eps,co)
 
-  if (co == 0) then
-#if (ABV == 0)  
-    call ricci_gamma(ex, X, Y, Z,                                      &
-                 chi,                                                  &
-                 dxx    ,   gxy    ,   gxz    ,   dyy    ,   gyz    ,   dzz,&
-                 Gamx   ,  Gamy    ,  Gamz    , &
-                 Gamxxx,Gamxxy,Gamxxz,Gamxyy,Gamxyz,Gamxzz,&
-                 Gamyxx,Gamyxy,Gamyxz,Gamyyy,Gamyyz,Gamyzz,&
-                 Gamzxx,Gamzxy,Gamzxz,Gamzyy,Gamzyz,Gamzzz,&
-                 Rxx,Rxy,Rxz,Ryy,Ryz,Rzz,&
-                 Symmetry)
-#endif
-    call constraint_bssn(ex, X, Y, Z,&
-                 chi,trK, &
-                 dxx,gxy,gxz,dyy,gyz,dzz, &
-                 Axx,Axy,Axz,Ayy,Ayz,Azz, &
-                 Gamx,Gamy,Gamz,&
-                 Lap,betax,betay,betaz,rho,Sx,Sy,Sz,&
-                 Gamxxx, Gamxxy, Gamxxz,Gamxyy, Gamxyz, Gamxzz, &
-                 Gamyxx, Gamyxy, Gamyxz,Gamyyy, Gamyyz, Gamyzz, &
-                 Gamzxx, Gamzxy, Gamzxz,Gamzyy, Gamzyz, Gamzzz, &
-                 Rxx,Rxy,Rxz,Ryy,Ryz,Rzz, &
-                 Hcon,Mxcon,Mycon,Mzcon,Gmxcon,Gmycon,Gmzcon, &
-                 Symmetry)
-  endif
+#if (ABV == 0)  
+  call ricci_gamma(ex, X, Y, Z,                                      &
+               chi,                                                  &
+               dxx    ,   gxy    ,   gxz    ,   dyy    ,   gyz    ,   dzz,&
+               Gamx   ,  Gamy    ,  Gamz    , &
+               Gamxxx,Gamxxy,Gamxxz,Gamxyy,Gamxyz,Gamxzz,&
+               Gamyxx,Gamyxy,Gamyxz,Gamyyy,Gamyyz,Gamyzz,&
+               Gamzxx,Gamzxy,Gamzxz,Gamzyy,Gamzyz,Gamzzz,&
+               Rxx,Rxy,Rxz,Ryy,Ryz,Rzz,&
+               Symmetry)
+#endif
+  call constraint_bssn(ex, X, Y, Z,&
+               chi,trK, &
+               dxx,gxy,gxz,dyy,gyz,dzz, &
+               Axx,Axy,Axz,Ayy,Ayz,Azz, &
+               Gamx,Gamy,Gamz,&
+               Lap,betax,betay,betaz,rho,Sx,Sy,Sz,&
+               Gamxxx, Gamxxy, Gamxxz,Gamxyy, Gamxyz, Gamxzz, &
+               Gamyxx, Gamyxy, Gamyxz,Gamyyy, Gamyyz, Gamyzz, &
+               Gamzxx, Gamzxy, Gamzxz,Gamzyy, Gamzyz, Gamzzz, &
+               Rxx,Rxy,Rxz,Ryy,Ryz,Rzz, &
+               Hcon,Mxcon,Mycon,Mzcon,Gmxcon,Gmycon,Gmzcon, &
+               Symmetry)
 
   return
 
@@ -228,12 +226,11 @@
 
   call get_Z4cparameters(kappa1,kappa2,kappa3,FF,eta)
 
-!!! sanity check
-#ifdef DEBUG
-  dX = sum(chi)+sum(trK)+sum(dxx)+sum(gxy)+sum(gxz)+sum(dyy)+sum(gyz)+sum(dzz) &
-      +sum(Axx)+sum(Axy)+sum(Axz)+sum(Ayy)+sum(Ayz)+sum(Azz)                   &
-      +sum(Gamx)+sum(Gamy)+sum(Gamz)                                           &
-      +sum(Lap)+sum(betax)+sum(betay)+sum(betaz)+sum(dtSfx)+sum(dtSfy)+sum(dtSfz) &
+!!! sanity check
+  dX = sum(chi)+sum(trK)+sum(dxx)+sum(gxy)+sum(gxz)+sum(dyy)+sum(gyz)+sum(dzz) &
+      +sum(Axx)+sum(Axy)+sum(Axz)+sum(Ayy)+sum(Ayz)+sum(Azz)                   &
+      +sum(Gamx)+sum(Gamy)+sum(Gamz)                                           &
+      +sum(Lap)+sum(betax)+sum(betay)+sum(betaz)+sum(dtSfx)+sum(dtSfy)+sum(dtSfz) &
       +sum(TZ)
   if(dX.ne.dX) then
      if(sum(chi).ne.sum(chi))write(*,*)"Z4c_rhs.f90: find NaN in chi"
@@ -260,11 +257,10 @@
      if(sum(dtSfx).ne.sum(dtSfx))write(*,*)"Z4c_rhs.f90: find NaN in dtSfx"
      if(sum(dtSfy).ne.sum(dtSfy))write(*,*)"Z4c_rhs.f90: find NaN in dtSfy"
      if(sum(dtSfz).ne.sum(dtSfz))write(*,*)"Z4c_rhs.f90: find NaN in dtSfz"
-     if(sum(TZ).ne.sum(Tz))write(*,*)"Z4c_rhs.f90: find NaN in TZ"
-     gont = 1
-     return
-  endif
-#endif
+     if(sum(TZ).ne.sum(Tz))write(*,*)"Z4c_rhs.f90: find NaN in TZ"
+     gont = 1
+     return
+  endif
 
   PI = dacos(-ONE)
 
@@ -1267,32 +1263,30 @@
 
   endif
 
-  if (co == 0) then
-#if (ABV == 0)  
-    call ricci_gamma(ex, X, Y, Z,                                      &
-                 chi,                                                  &
-                 dxx    ,   gxy    ,   gxz    ,   dyy    ,   gyz    ,   dzz,&
-                 Gamx   ,  Gamy    ,  Gamz    , &
-                 Gamxxx,Gamxxy,Gamxxz,Gamxyy,Gamxyz,Gamxzz,&
-                 Gamyxx,Gamyxy,Gamyxz,Gamyyy,Gamyyz,Gamyzz,&
-                 Gamzxx,Gamzxy,Gamzxz,Gamzyy,Gamzyz,Gamzzz,&
-                 Rxx,Rxy,Rxz,Ryy,Ryz,Rzz,&
-                 Symmetry)
-#endif
-
-    call constraint_bssn(ex, X, Y, Z,&
-                 chi,trK, &
-                 dxx,gxy,gxz,dyy,gyz,dzz, &
-                 Axx,Axy,Axz,Ayy,Ayz,Azz, &
-                 Gamx,Gamy,Gamz,&
-                 Lap,betax,betay,betaz,rho,Sx,Sy,Sz,&
-                 Gamxxx, Gamxxy, Gamxxz,Gamxyy, Gamxyz, Gamxzz, &
-                 Gamyxx, Gamyxy, Gamyxz,Gamyyy, Gamyyz, Gamyzz, &
-                 Gamzxx, Gamzxy, Gamzxz,Gamzyy, Gamzyz, Gamzzz, &
-                 Rxx,Rxy,Rxz,Ryy,Ryz,Rzz, &
-                 Hcon,Mxcon,Mycon,Mzcon,Gmxcon,Gmycon,Gmzcon, &
-                 Symmetry)
-  endif
+#if (ABV == 0)  
+  call ricci_gamma(ex, X, Y, Z,                                      &
+               chi,                                                  &
+               dxx    ,   gxy    ,   gxz    ,   dyy    ,   gyz    ,   dzz,&
+               Gamx   ,  Gamy    ,  Gamz    , &
+               Gamxxx,Gamxxy,Gamxxz,Gamxyy,Gamxyz,Gamxzz,&
+               Gamyxx,Gamyxy,Gamyxz,Gamyyy,Gamyyz,Gamyzz,&
+               Gamzxx,Gamzxy,Gamzxz,Gamzyy,Gamzyz,Gamzzz,&
+               Rxx,Rxy,Rxz,Ryy,Ryz,Rzz,&
+               Symmetry)
+#endif
+
+  call constraint_bssn(ex, X, Y, Z,&
+               chi,trK, &
+               dxx,gxy,gxz,dyy,gyz,dzz, &
+               Axx,Axy,Axz,Ayy,Ayz,Azz, &
+               Gamx,Gamy,Gamz,&
+               Lap,betax,betay,betaz,rho,Sx,Sy,Sz,&
+               Gamxxx, Gamxxy, Gamxxz,Gamxyy, Gamxyz, Gamxzz, &
+               Gamyxx, Gamyxy, Gamyxz,Gamyyy, Gamyyz, Gamyzz, &
+               Gamzxx, Gamzxy, Gamzxz,Gamzyy, Gamzyz, Gamzzz, &
+               Rxx,Rxy,Rxz,Ryy,Ryz,Rzz, &
+               Hcon,Mxcon,Mycon,Mzcon,Gmxcon,Gmycon,Gmzcon, &
+               Symmetry)
 
   gont = 0
 

AMSS_NCKU_source/Z4c_rhs_ss.f90

@@ -121,12 +121,11 @@
 
   call get_Z4cparameters(kappa1,kappa2,kappa3,FF,eta)
 
-!!! sanity check
-#ifdef DEBUG
-  dX = sum(chi)+sum(trK)+sum(dxx)+sum(gxy)+sum(gxz)+sum(dyy)+sum(gyz)+sum(dzz) &
-      +sum(Axx)+sum(Axy)+sum(Axz)+sum(Ayy)+sum(Ayz)+sum(Azz)                   &
-      +sum(Gamx)+sum(Gamy)+sum(Gamz)                                           &
-      +sum(Lap)+sum(betax)+sum(betay)+sum(betaz)+sum(dtSfx)+sum(dtSfy)+sum(dtSfz) &
+!!! sanity check
+  dX = sum(chi)+sum(trK)+sum(dxx)+sum(gxy)+sum(gxz)+sum(dyy)+sum(gyz)+sum(dzz) &
+      +sum(Axx)+sum(Axy)+sum(Axz)+sum(Ayy)+sum(Ayz)+sum(Azz)                   &
+      +sum(Gamx)+sum(Gamy)+sum(Gamz)                                           &
+      +sum(Lap)+sum(betax)+sum(betay)+sum(betaz)+sum(dtSfx)+sum(dtSfy)+sum(dtSfz) &
       +sum(TZ)
   if(dX.ne.dX) then
      if(sum(chi).ne.sum(chi))write(*,*)"Z4c_rhs_ss.f90: find NaN in chi"
@@ -153,11 +152,10 @@
      if(sum(dtSfx).ne.sum(dtSfx))write(*,*)"Z4c_rhs_ss.f90: find NaN in dtSfx"
      if(sum(dtSfy).ne.sum(dtSfy))write(*,*)"Z4c_rhs_ss.f90: find NaN in dtSfy"
      if(sum(dtSfz).ne.sum(dtSfz))write(*,*)"Z4c_rhs_ss.f90: find NaN in dtSfz"
-     if(sum(TZ).ne.sum(Tz))write(*,*)"Z4c_rhs_ss.f90: find NaN in TZ"
-     gont = 1
-     return
-  endif
-#endif
+     if(sum(TZ).ne.sum(Tz))write(*,*)"Z4c_rhs_ss.f90: find NaN in TZ"
+     gont = 1
+     return
+  endif
 
   PI = dacos(-ONE)
 
@@ -1390,43 +1388,41 @@
   call kodis_sh(ex,crho,sigma,R,TZ,TZ_rhs,SSS,Symmetry,eps,sst)
   endif
 
-  if (co == 0) then
-#if (ABV == 1)  
-    call ricci_gamma_ss(ex,crho,sigma,R,X, Y, Z,                               &
-                 drhodx, drhody, drhodz,                                       &
-                 dsigmadx,dsigmady,dsigmadz,                                   &
-                 dRdx,dRdy,dRdz,                                               &
-                 drhodxx,drhodxy,drhodxz,drhodyy,drhodyz,drhodzz,              &
-                 dsigmadxx,dsigmadxy,dsigmadxz,dsigmadyy,dsigmadyz,dsigmadzz,  &
-                 dRdxx,dRdxy,dRdxz,dRdyy,dRdyz,dRdzz,                          &
-                 chi,                                                          &
-                 dxx    ,   gxy    ,   gxz    ,   dyy    ,   gyz    ,   dzz,&
-                 Gamx   ,  Gamy    ,  Gamz    , &
-                 Gamxxx,Gamxxy,Gamxxz,Gamxyy,Gamxyz,Gamxzz,&
-                 Gamyxx,Gamyxy,Gamyxz,Gamyyy,Gamyyz,Gamyzz,&
-                 Gamzxx,Gamzxy,Gamzxz,Gamzyy,Gamzyz,Gamzzz,&
-                 Rxx,Rxy,Rxz,Ryy,Ryz,Rzz,&
-                 Symmetry,Lev,sst)
-#endif
-    call constraint_bssn_ss(ex,crho,sigma,R,X, Y, Z,  &
-                 drhodx, drhody, drhodz,                                       &
-                 dsigmadx,dsigmady,dsigmadz,                                   &
-                 dRdx,dRdy,dRdz,                                               &
-                 drhodxx,drhodxy,drhodxz,drhodyy,drhodyz,drhodzz,              &
-                 dsigmadxx,dsigmadxy,dsigmadxz,dsigmadyy,dsigmadyz,dsigmadzz,  &
-                 dRdxx,dRdxy,dRdxz,dRdyy,dRdyz,dRdzz,                          &
-                 chi,trK, &
-                 dxx,gxy,gxz,dyy,gyz,dzz, &
-                 Axx,Axy,Axz,Ayy,Ayz,Azz, &
-                 Gamx,Gamy,Gamz,&
-                 Lap,betax,betay,betaz,rho,Sx,Sy,Sz,&
-                 Gamxxx, Gamxxy, Gamxxz,Gamxyy, Gamxyz, Gamxzz, &
-                 Gamyxx, Gamyxy, Gamyxz,Gamyyy, Gamyyz, Gamyzz, &
-                 Gamzxx, Gamzxy, Gamzxz,Gamzyy, Gamzyz, Gamzzz, &
-                 Rxx,Rxy,Rxz,Ryy,Ryz,Rzz, &
-                 Hcon,Mxcon,Mycon,Mzcon,Gmxcon,Gmycon,Gmzcon, &
-                 Symmetry,Lev,sst)
-  endif
+#if (ABV == 1)  
+  call ricci_gamma_ss(ex,crho,sigma,R,X, Y, Z,                                 &
+               drhodx, drhody, drhodz,                                         &
+               dsigmadx,dsigmady,dsigmadz,                                     &
+               dRdx,dRdy,dRdz,                                                 &
+               drhodxx,drhodxy,drhodxz,drhodyy,drhodyz,drhodzz,                &
+               dsigmadxx,dsigmadxy,dsigmadxz,dsigmadyy,dsigmadyz,dsigmadzz,    &
+               dRdxx,dRdxy,dRdxz,dRdyy,dRdyz,dRdzz,                            &
+               chi,                                                  &
+               dxx    ,   gxy    ,   gxz    ,   dyy    ,   gyz    ,   dzz,&
+               Gamx   ,  Gamy    ,  Gamz    , &
+               Gamxxx,Gamxxy,Gamxxz,Gamxyy,Gamxyz,Gamxzz,&
+               Gamyxx,Gamyxy,Gamyxz,Gamyyy,Gamyyz,Gamyzz,&
+               Gamzxx,Gamzxy,Gamzxz,Gamzyy,Gamzyz,Gamzzz,&
+               Rxx,Rxy,Rxz,Ryy,Ryz,Rzz,&
+               Symmetry,Lev,sst)
+  call constraint_bssn_ss(ex,crho,sigma,R,X, Y, Z,  &
+               drhodx, drhody, drhodz,                                         &
+               dsigmadx,dsigmady,dsigmadz,                                     &
+               dRdx,dRdy,dRdz,                                                 &
+               drhodxx,drhodxy,drhodxz,drhodyy,drhodyz,drhodzz,                &
+               dsigmadxx,dsigmadxy,dsigmadxz,dsigmadyy,dsigmadyz,dsigmadzz,    &
+               dRdxx,dRdxy,dRdxz,dRdyy,dRdyz,dRdzz,                            &
+               chi,trK, &
+               dxx,gxy,gxz,dyy,gyz,dzz, &
+               Axx,Axy,Axz,Ayy,Ayz,Azz, &
+               Gamx,Gamy,Gamz,&
+               Lap,betax,betay,betaz,rho,Sx,Sy,Sz,&
+               Gamxxx, Gamxxy, Gamxxz,Gamxyy, Gamxyz, Gamxzz, &
+               Gamyxx, Gamyxy, Gamyxz,Gamyyy, Gamyyz, Gamyzz, &
+               Gamzxx, Gamzxy, Gamzxz,Gamzyy, Gamzyz, Gamzzz, &
+               Rxx,Rxy,Rxz,Ryy,Ryz,Rzz, &
+               Hcon,Mxcon,Mycon,Mzcon,Gmxcon,Gmycon,Gmzcon, &
+               Symmetry,Lev,sst)
+#endif
 
   gont = 0
 

AMSS_NCKU_source/bssnEM_class.C

@@ -258,8 +258,6 @@ void bssnEM_class::Initialize()
     PhysTime = StartTime;
     Setup_Black_Hole_position();
   }
-
-  setup_transfer_caches();
 }
 
 //================================================================================================

AMSS_NCKU_source/bssnEScalar_class.C

@@ -23,14 +23,8 @@ using namespace std;
 #include "rungekutta4_rout.h"
 #include "sommerfeld_rout.h"
 #include "getnp4.h"
-#include "shellfunctions.h"
-#include "parameters.h"
-
-#if BSSN_USE_ESCALAR_C_KERNEL
-#define BSSN_ESCALAR_RHS f_compute_rhs_bssn_escalar_c
-#else
-#define BSSN_ESCALAR_RHS f_compute_rhs_bssn_escalar
-#endif
+#include "shellfunctions.h"
+#include "parameters.h"
 
 #ifdef With_AHF
 #include "derivatives.h"
@@ -80,8 +74,8 @@ bssnEScalar_class::bssnEScalar_class(double Couranti, double StartTimei, double
 
 //================================================================================================
 
-void bssnEScalar_class::Initialize()
-{
+void bssnEScalar_class::Initialize()
+{
   Sphio = new var("Sphio", ngfs++, 1, 1, 1);
   Spio = new var("Spio", ngfs++, 1, 1, 1);
   Sphi0 = new var("Sphi0", ngfs++, 1, 1, 1);
@@ -138,14 +132,11 @@ void bssnEScalar_class::Initialize()
       }
     } 
    
-  GH = new cgh(0, ngfs, Symmetry, pname, checkrun, ErrorMonitor);
-  ConstraintRefreshLevels = new int[GH->levels];
-  for (int il = 0; il < GH->levels; il++)
-    ConstraintRefreshLevels[il] = 0;
-  if (checkrun)
-    CheckPoint->readcheck_cgh(PhysTime, GH, myrank, nprocs, Symmetry);
-  else
-    GH->compose_cgh(nprocs);
+  GH = new cgh(0, ngfs, Symmetry, pname, checkrun, ErrorMonitor);
+  if (checkrun)
+    CheckPoint->readcheck_cgh(PhysTime, GH, myrank, nprocs, Symmetry);
+  else
+    GH->compose_cgh(nprocs);
 
 #ifdef WithShell
   SH = new ShellPatch(0, ngfs, pname, Symmetry, myrank, ErrorMonitor);
@@ -169,14 +160,12 @@ void bssnEScalar_class::Initialize()
   {
     CheckPoint->read_Black_Hole_position(BH_num_input, BH_num, Porg0, Pmom, Spin, Mass, Porgbr, Porg, Porg1, Porg_rhs);
   }
-  else
-  {
-    PhysTime = StartTime;
-    Setup_Black_Hole_position();
-  }
-
-  setup_transfer_caches();
-}
+  else
+  {
+    PhysTime = StartTime;
+    Setup_Black_Hole_position();
+  }
+}
 
 //================================================================================================
 
@@ -218,10 +207,10 @@ bssnEScalar_class::~bssnEScalar_class()
 
 // Read initial data solved by Ansorg, PRD 70, 064011 (2004)
 
-void bssnEScalar_class::Read_Ansorg()
-{
-  if (!checkrun)
-  {
+void bssnEScalar_class::Read_Ansorg()
+{
+  if (!checkrun)
+  {
     if (myrank == 0)
       cout << "Read initial data from Ansorg's solver,"
            << " please be sure the input parameters for black holes are puncture parameters!!" 
@@ -238,12 +227,9 @@ void bssnEScalar_class::Read_Ansorg()
         cout << "Error inputpar" << endl;
         exit(0);
       }
-    }
-    int BH_NM;
-    double *Porg_here;
-    double *pmom_local;
-    double *spin_local;
-    double *mass_local;
+    }
+    int BH_NM;
+    double *Porg_here;
     // read parameter from file
     {
       const int LEN = 256;
@@ -283,11 +269,11 @@ void bssnEScalar_class::Read_Ansorg()
       }
       inf.close();
     }
-
-    Porg_here = new double[3 * BH_NM];
-    pmom_local = new double[3 * BH_NM];
-    spin_local = new double[3 * BH_NM];
-    mass_local = new double[BH_NM];
+
+    Porg_here = new double[3 * BH_NM];
+    Pmom = new double[3 * BH_NM];
+    Spin = new double[3 * BH_NM];
+    Mass = new double[BH_NM];
     // read parameter from file
     {
       const int LEN = 256;
@@ -319,37 +305,37 @@ void bssnEScalar_class::Read_Ansorg()
         else if (status == 0)
           continue;
 
-        if (sgrp == "BSSN" && sind < BH_NM)
-        {
-          if (skey == "Mass")
-            mass_local[sind] = atof(sval.c_str());
-          else if (skey == "Porgx")
-            Porg_here[sind * 3] = atof(sval.c_str());
-          else if (skey == "Porgy")
-            Porg_here[sind * 3 + 1] = atof(sval.c_str());
-          else if (skey == "Porgz")
-            Porg_here[sind * 3 + 2] = atof(sval.c_str());
-          else if (skey == "Spinx")
-            spin_local[sind * 3] = atof(sval.c_str());
-          else if (skey == "Spiny")
-            spin_local[sind * 3 + 1] = atof(sval.c_str());
-          else if (skey == "Spinz")
-            spin_local[sind * 3 + 2] = atof(sval.c_str());
-          else if (skey == "Pmomx")
-            pmom_local[sind * 3] = atof(sval.c_str());
-          else if (skey == "Pmomy")
-            pmom_local[sind * 3 + 1] = atof(sval.c_str());
-          else if (skey == "Pmomz")
-            pmom_local[sind * 3 + 2] = atof(sval.c_str());
-        }
-      }
-      inf.close();
+        if (sgrp == "BSSN" && sind < BH_NM)
+        {
+          if (skey == "Mass")
+            Mass[sind] = atof(sval.c_str());
+          else if (skey == "Porgx")
+            Porg_here[sind * 3] = atof(sval.c_str());
+          else if (skey == "Porgy")
+            Porg_here[sind * 3 + 1] = atof(sval.c_str());
+          else if (skey == "Porgz")
+            Porg_here[sind * 3 + 2] = atof(sval.c_str());
+          else if (skey == "Spinx")
+            Spin[sind * 3] = atof(sval.c_str());
+          else if (skey == "Spiny")
+            Spin[sind * 3 + 1] = atof(sval.c_str());
+          else if (skey == "Spinz")
+            Spin[sind * 3 + 2] = atof(sval.c_str());
+          else if (skey == "Pmomx")
+            Pmom[sind * 3] = atof(sval.c_str());
+          else if (skey == "Pmomy")
+            Pmom[sind * 3 + 1] = atof(sval.c_str());
+          else if (skey == "Pmomz")
+            Pmom[sind * 3 + 2] = atof(sval.c_str());
+        }
+      }
+      inf.close();
     }
-    int order = 6;
-    Ansorg read_ansorg("Ansorg.psid", order);
-    // set initial data
-    for (int lev = 0; lev < GH->levels; lev++)
-    {
+    int order = 6;
+    Ansorg read_ansorg("Ansorg.psid", order);
+    // set initial data
+    for (int lev = 0; lev < GH->levels; lev++)
+    {
       MyList<Patch> *Pp = GH->PatL[lev];
       while (Pp)
       {
@@ -372,21 +358,21 @@ void bssnEScalar_class::Read_Ansorg()
                                       cg->fgfs[Axx0->sgfn], cg->fgfs[Axy0->sgfn], cg->fgfs[Axz0->sgfn], 
                                       cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn],
                                       cg->fgfs[Gmx0->sgfn], cg->fgfs[Gmy0->sgfn], cg->fgfs[Gmz0->sgfn],
-                                      cg->fgfs[Lap0->sgfn], 
-                                      cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
-                                      cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
-                                      cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
-                                      mass_local, Porg_here, pmom_local, spin_local, BH_NM);
+                                      cg->fgfs[Lap0->sgfn], 
+                                      cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
+                                      cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
+                                      cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
+                                      Mass, Porg_here, Pmom, Spin, BH_NM);
           }
           if (BL == Pp->data->ble)
             break;
           BL = BL->next;
-        }
-        Pp = Pp->next;
-      }
-    }
-#ifdef WithShell
-    // ShellPatch part
+        }
+        Pp = Pp->next;
+      }
+    }
+#ifdef WithShell
+    // ShellPatch part
     MyList<ss_patch> *Pp = SH->PatL;
     while (Pp)
     {
@@ -414,28 +400,25 @@ void bssnEScalar_class::Read_Ansorg()
                                        cg->fgfs[Axx0->sgfn], cg->fgfs[Axy0->sgfn], cg->fgfs[Axz0->sgfn], 
                                        cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn],
                                        cg->fgfs[Gmx0->sgfn], cg->fgfs[Gmy0->sgfn], cg->fgfs[Gmz0->sgfn],
-                                       cg->fgfs[Lap0->sgfn], 
-                                       cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
-                                       cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
-                                       cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
-                                       mass_local, Porg_here, pmom_local, spin_local, BH_NM);
+                                       cg->fgfs[Lap0->sgfn], 
+                                       cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
+                                       cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
+                                       cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
+                                       Mass, Porg_here, Pmom, Spin, BH_NM);
         }
         if (BL == Pp->data->ble)
           break;
         BL = BL->next;
-      }
-      Pp = Pp->next;
-    }
-#endif
-
-    delete[] Porg_here;
-    delete[] pmom_local;
-    delete[] spin_local;
-    delete[] mass_local;
-    // dump read_in initial data
-    //   for(int lev=0;lev<GH->levels;lev++) Parallel::Dump_Data(GH->PatL[lev],StateList,0,PhysTime,dT);
-  }
-}
+      }
+      Pp = Pp->next;
+    }
+#endif
+
+    delete[] Porg_here;
+    // dump read_in initial data
+    //   for(int lev=0;lev<GH->levels;lev++) Parallel::Dump_Data(GH->PatL[lev],StateList,0,PhysTime,dT);
+  }
+}
 
 //================================================================================================
 
@@ -449,10 +432,10 @@ void bssnEScalar_class::Read_Ansorg()
 
 // Read initial data solved by Pablo's Olliptic Phys.Rev.D 82 024005 (2010)
 
-void bssnEScalar_class::Read_Pablo()
-{
-  if (!checkrun)
-  {
+void bssnEScalar_class::Read_Pablo()
+{
+  if (!checkrun)
+  {
     if (myrank == 0)
       cout << "Read initial data from Pablo's solver,"
            << " please be sure the input parameters for black holes are puncture parameters!!" 
@@ -469,12 +452,9 @@ void bssnEScalar_class::Read_Pablo()
         cout << "Error inputpar" << endl;
         exit(0);
       }
-    }
-    int BH_NM;
-    double *Porg_here;
-    double *pmom_local;
-    double *spin_local;
-    double *mass_local;
+    }
+    int BH_NM;
+    double *Porg_here;
     // read parameter from file
     {
       const int LEN = 256;
@@ -514,11 +494,11 @@ void bssnEScalar_class::Read_Pablo()
       }
       inf.close();
     }
-
-    Porg_here = new double[3 * BH_NM];
-    pmom_local = new double[3 * BH_NM];
-    spin_local = new double[3 * BH_NM];
-    mass_local = new double[BH_NM];
+
+    Porg_here = new double[3 * BH_NM];
+    Pmom = new double[3 * BH_NM];
+    Spin = new double[3 * BH_NM];
+    Mass = new double[BH_NM];
     // read parameter from file
     {
       const int LEN = 256;
@@ -550,31 +530,31 @@ void bssnEScalar_class::Read_Pablo()
         else if (status == 0)
           continue;
 
-        if (sgrp == "BSSN" && sind < BH_NM)
-        {
-          if (skey == "Mass")
-            mass_local[sind] = atof(sval.c_str());
-          else if (skey == "Porgx")
-            Porg_here[sind * 3] = atof(sval.c_str());
-          else if (skey == "Porgy")
-            Porg_here[sind * 3 + 1] = atof(sval.c_str());
-          else if (skey == "Porgz")
-            Porg_here[sind * 3 + 2] = atof(sval.c_str());
-          else if (skey == "Spinx")
-            spin_local[sind * 3] = atof(sval.c_str());
-          else if (skey == "Spiny")
-            spin_local[sind * 3 + 1] = atof(sval.c_str());
-          else if (skey == "Spinz")
-            spin_local[sind * 3 + 2] = atof(sval.c_str());
-          else if (skey == "Pmomx")
-            pmom_local[sind * 3] = atof(sval.c_str());
-          else if (skey == "Pmomy")
-            pmom_local[sind * 3 + 1] = atof(sval.c_str());
-          else if (skey == "Pmomz")
-            pmom_local[sind * 3 + 2] = atof(sval.c_str());
-        }
-      }
-      inf.close();
+        if (sgrp == "BSSN" && sind < BH_NM)
+        {
+          if (skey == "Mass")
+            Mass[sind] = atof(sval.c_str());
+          else if (skey == "Porgx")
+            Porg_here[sind * 3] = atof(sval.c_str());
+          else if (skey == "Porgy")
+            Porg_here[sind * 3 + 1] = atof(sval.c_str());
+          else if (skey == "Porgz")
+            Porg_here[sind * 3 + 2] = atof(sval.c_str());
+          else if (skey == "Spinx")
+            Spin[sind * 3] = atof(sval.c_str());
+          else if (skey == "Spiny")
+            Spin[sind * 3 + 1] = atof(sval.c_str());
+          else if (skey == "Spinz")
+            Spin[sind * 3 + 2] = atof(sval.c_str());
+          else if (skey == "Pmomx")
+            Pmom[sind * 3] = atof(sval.c_str());
+          else if (skey == "Pmomy")
+            Pmom[sind * 3 + 1] = atof(sval.c_str());
+          else if (skey == "Pmomz")
+            Pmom[sind * 3 + 2] = atof(sval.c_str());
+        }
+      }
+      inf.close();
     }
     bool flag = false;
     int DIM = dim;
@@ -614,11 +594,11 @@ void bssnEScalar_class::Read_Pablo()
                                         cg->fgfs[Axx0->sgfn], cg->fgfs[Axy0->sgfn], cg->fgfs[Axz0->sgfn], 
                                         cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn],
                                         cg->fgfs[Gmx0->sgfn], cg->fgfs[Gmy0->sgfn], cg->fgfs[Gmz0->sgfn],
-                                        cg->fgfs[Lap0->sgfn], 
-                                        cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
-                                        cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
-                                        cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
-                                        mass_local, Porg_here, pmom_local, spin_local, BH_NM);
+                                        cg->fgfs[Lap0->sgfn], 
+                                        cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
+                                        cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
+                                        cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
+                                        Mass, Porg_here, Pmom, Spin, BH_NM);
             }
             if (BL == Pp->data->ble)
               break;
@@ -678,11 +658,11 @@ void bssnEScalar_class::Read_Pablo()
                                          cg->fgfs[Axx0->sgfn], cg->fgfs[Axy0->sgfn], cg->fgfs[Axz0->sgfn], 
                                          cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn],
                                          cg->fgfs[Gmx0->sgfn], cg->fgfs[Gmy0->sgfn], cg->fgfs[Gmz0->sgfn],
-                                         cg->fgfs[Lap0->sgfn], 
-                                         cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
-                                         cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
-                                         cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
-                                         mass_local, Porg_here, pmom_local, spin_local, BH_NM);
+                                         cg->fgfs[Lap0->sgfn], 
+                                         cg->fgfs[Sfx0->sgfn], cg->fgfs[Sfy0->sgfn], cg->fgfs[Sfz0->sgfn],
+                                         cg->fgfs[dtSfx0->sgfn], cg->fgfs[dtSfy0->sgfn], cg->fgfs[dtSfz0->sgfn],
+                                         cg->fgfs[Sphi0->sgfn], cg->fgfs[Spi0->sgfn],
+                                         Mass, Porg_here, Pmom, Spin, BH_NM);
           }
           if (BL == Pp->data->ble)
             break;
@@ -704,13 +684,10 @@ void bssnEScalar_class::Read_Pablo()
       Pp = Pp->next;
     }
 #endif
-
-    delete[] Porg_here;
-    delete[] pmom_local;
-    delete[] spin_local;
-    delete[] mass_local;
-    if (flag && myrank == 0)
-      MPI_Abort(MPI_COMM_WORLD, 1);
+
+    delete[] Porg_here;
+    if (flag && myrank == 0)
+      MPI_Abort(MPI_COMM_WORLD, 1);
     // dump read_in initial data
     for (int lev = 0; lev < GH->levels; lev++)
       Parallel::Dump_Data(GH->PatL[lev], StateList, 0, PhysTime, dT);
@@ -762,10 +739,10 @@ void bssnEScalar_class::Step(int lev, int YN)
                      cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn]);
 #endif
 
-        if (BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
-                                       cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
-                                       cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn], 
-                                       cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
+        if (f_compute_rhs_bssn_escalar(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
+                                       cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
+                                       cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn], 
+                                       cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
                                        cg->fgfs[Axx0->sgfn], cg->fgfs[Axy0->sgfn], cg->fgfs[Axz0->sgfn], 
                                        cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn],
                                        cg->fgfs[Gmx0->sgfn], cg->fgfs[Gmy0->sgfn], cg->fgfs[Gmz0->sgfn],
@@ -1016,12 +993,11 @@ void bssnEScalar_class::Step(int lev, int YN)
   }
 #endif
 
-  Parallel::AsyncSyncState async_pre;
-  sync_predictor_start(lev, SynchList_pre, async_pre);
+  Parallel::Sync(GH->PatL[lev], SynchList_pre, Symmetry);
 
-#ifdef WithShell
-  if (lev == 0)
-  {
+#ifdef WithShell
+  if (lev == 0)
+  {
     clock_t prev_clock, curr_clock;
     if (myrank == 0)
       curr_clock = clock();
@@ -1033,10 +1009,9 @@ void bssnEScalar_class::Step(int lev, int YN)
       cout << " Shell stuff synchronization used " 
            << (double)(curr_clock - prev_clock) / ((double)CLOCKS_PER_SEC) 
            << " seconds! " << endl;
-    }
-  }
-#endif
-  sync_predictor_finish(lev, async_pre, SynchList_pre);
+    }
+  }
+#endif
 
   // for black hole position
   if (BH_num > 0 && lev == GH->levels - 1)
@@ -1106,10 +1081,10 @@ void bssnEScalar_class::Step(int lev, int YN)
                          cg->fgfs[Ayy->sgfn], cg->fgfs[Ayz->sgfn], cg->fgfs[Azz->sgfn]);
 #endif
 
-          if (BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
-                                         cg->fgfs[phi->sgfn], cg->fgfs[trK->sgfn],
-                                         cg->fgfs[gxx->sgfn], cg->fgfs[gxy->sgfn], cg->fgfs[gxz->sgfn], 
-                                         cg->fgfs[gyy->sgfn], cg->fgfs[gyz->sgfn], cg->fgfs[gzz->sgfn],
+          if (f_compute_rhs_bssn_escalar(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
+                                         cg->fgfs[phi->sgfn], cg->fgfs[trK->sgfn],
+                                         cg->fgfs[gxx->sgfn], cg->fgfs[gxy->sgfn], cg->fgfs[gxz->sgfn], 
+                                         cg->fgfs[gyy->sgfn], cg->fgfs[gyz->sgfn], cg->fgfs[gzz->sgfn],
                                          cg->fgfs[Axx->sgfn], cg->fgfs[Axy->sgfn], cg->fgfs[Axz->sgfn], 
                                          cg->fgfs[Ayy->sgfn], cg->fgfs[Ayz->sgfn], cg->fgfs[Azz->sgfn],
                                          cg->fgfs[Gmx->sgfn], cg->fgfs[Gmy->sgfn], cg->fgfs[Gmz->sgfn],
@@ -1374,12 +1349,11 @@ void bssnEScalar_class::Step(int lev, int YN)
     }
 #endif
 
-    Parallel::AsyncSyncState async_cor;
-    sync_corrector_start(lev, SynchList_cor, async_cor);
+    Parallel::Sync(GH->PatL[lev], SynchList_cor, Symmetry);
 
-#ifdef WithShell
-    if (lev == 0)
-    {
+#ifdef WithShell
+    if (lev == 0)
+    {
       clock_t prev_clock, curr_clock;
       if (myrank == 0)
         curr_clock = clock();
@@ -1391,10 +1365,9 @@ void bssnEScalar_class::Step(int lev, int YN)
         cout << " Shell stuff synchronization used " 
              << (double)(curr_clock - prev_clock) / ((double)CLOCKS_PER_SEC) 
              << " seconds! " << endl;
-      }
-    }
-#endif
-    sync_corrector_finish(lev, async_cor, SynchList_cor);
+      }
+    }
+#endif
     // for black hole position
     if (BH_num > 0 && lev == GH->levels - 1)
     {
@@ -1862,14 +1835,11 @@ void bssnEScalar_class::AnalysisStuff_EScalar(int lev, double dT_lev)
 
 //================================================================================================
 
-void bssnEScalar_class::Interp_Constraint(bool infg)
-{
-  if (!infg)
-    return;
-
-  // we do not support a_lev != 0 yet.
-  if (a_lev > 0)
-    return;
+void bssnEScalar_class::Interp_Constraint()
+{
+  // we do not support a_lev != 0 yet.
+  if (a_lev > 0)
+    return;
 
   for (int lev = 0; lev < GH->levels; lev++)
   {
@@ -1888,10 +1858,10 @@ void bssnEScalar_class::Interp_Constraint(bool infg)
           if (myrank == cg->rank)
           {
             if (lev > 0)
-              BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
-                                         cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
-                                         cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn], 
-                                         cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
+              f_compute_rhs_bssn_escalar(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
+                                         cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
+                                         cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn], 
+                                         cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
                                          cg->fgfs[Axx0->sgfn], cg->fgfs[Axy0->sgfn], cg->fgfs[Axz0->sgfn], 
                                          cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn],
                                          cg->fgfs[Gmx0->sgfn], cg->fgfs[Gmy0->sgfn], cg->fgfs[Gmz0->sgfn],
@@ -2108,10 +2078,10 @@ void bssnEScalar_class::Constraint_Out()
             if (myrank == cg->rank)
             {
               if (lev > 0)
-                BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
-                                           cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
-                                           cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn], 
-                                           cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
+                f_compute_rhs_bssn_escalar(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
+                                           cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
+                                           cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn], 
+                                           cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
                                            cg->fgfs[Axx0->sgfn], cg->fgfs[Axy0->sgfn], cg->fgfs[Axz0->sgfn], 
                                            cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn],
                                            cg->fgfs[Gmx0->sgfn], cg->fgfs[Gmy0->sgfn], cg->fgfs[Gmz0->sgfn],

AMSS_NCKU_source/bssnEScalar_class.h

@@ -51,7 +51,7 @@ public:
      void Compute_Psi4(int lev);
      void Step(int lev, int YN);
      void AnalysisStuff_EScalar(int lev, double dT_lev);
-     void Interp_Constraint(bool infg);
+     void Interp_Constraint();
      void Constraint_Out(); 
 
 protected:

AMSS_NCKU_source/bssn_class.h

@@ -31,19 +31,11 @@ using namespace std;
 #include "surface_integral.h"
 #include "checkpoint.h"
 
-extern void setpbh(int iBHN, double **iPBH, double *iMass, int rBHN);
-
-#ifndef BSSN_USE_TRANSFER_CACHE
-#define BSSN_USE_TRANSFER_CACHE 1
-#endif
-
-#ifndef BSSN_USE_ESCALAR_C_KERNEL
-#define BSSN_USE_ESCALAR_C_KERNEL 1
-#endif
-
-class bssn_class
-{
-public:
+extern void setpbh(int iBHN, double **iPBH, double *iMass, int rBHN);
+
+class bssn_class
+{
+public:
        int ngfs;
        int nprocs, myrank;
        cgh *GH;
@@ -53,11 +45,10 @@ public:
        int checkrun;
        char checkfilename[50];
        int Steps;
-       double StartTime, TotalTime;
-       double AnasTime, DumpTime, d2DumpTime, CheckTime;
-       double LastAnas, LastConsOut;
-       int *ConstraintRefreshLevels;
-       double Courant;
+       double StartTime, TotalTime;
+       double AnasTime, DumpTime, d2DumpTime, CheckTime;
+       double LastAnas, LastConsOut;
+       double Courant;
        double numepss, numepsb, numepsh;
        int Symmetry;
        int maxl, decn;
@@ -137,14 +128,15 @@ public:
 
        Parallel::SyncCache *sync_cache_pre;  // per-level cache for predictor sync
        Parallel::SyncCache *sync_cache_cor;  // per-level cache for corrector sync
-       Parallel::SyncCache *sync_cache_rp_coarse;  // RestrictProlong sync on PatL[lev-1]
-       Parallel::SyncCache *sync_cache_rp_fine;    // RestrictProlong sync on PatL[lev]
-       Parallel::SyncCache *sync_cache_restrict;   // cached Restrict in RestrictProlong
-       Parallel::SyncCache *sync_cache_outbd;      // cached OutBdLow2Hi in RestrictProlong
+       Parallel::SyncCache *sync_cache_rp_coarse;  // RestrictProlong sync on PatL[lev-1]
+       Parallel::SyncCache *sync_cache_rp_fine;    // RestrictProlong sync on PatL[lev]
+       Parallel::SyncCache *sync_cache_restrict;   // cached Restrict in RestrictProlong
+       Parallel::SyncCache *sync_cache_outbd;      // cached OutBdLow2Hi in RestrictProlong
+       Parallel::SyncCache *sync_cache_psi4;       // cached Psi4 sync on PatL[lev]
 
-       monitor *ErrorMonitor, *Psi4Monitor, *BHMonitor, *MAPMonitor;
-       monitor *ConVMonitor, *TimingMonitor;
-       surface_integral *Waveshell;
+       monitor *ErrorMonitor, *Psi4Monitor, *BHMonitor, *MAPMonitor;
+       monitor *ConVMonitor;
+       surface_integral *Waveshell;
        checkpoint *CheckPoint;
 
 public:
@@ -175,32 +167,25 @@ public:
        void Setup_KerrSchild();
        void Enforce_algcon(int lev, int fg);
 
-       void testRestrict();
-       void testOutBd();
-       
-       bool check_Stdin_Abort(); 
-       bool use_transfer_cache() const;
-       void setup_transfer_caches();
-       void invalidate_transfer_caches();
-       void destroy_transfer_caches();
-       void sync_predictor_start(int lev, MyList<var> *VarList, Parallel::AsyncSyncState &async_state);
-       void sync_predictor_finish(int lev, Parallel::AsyncSyncState &async_state, MyList<var> *VarList);
-       void sync_corrector_start(int lev, MyList<var> *VarList, Parallel::AsyncSyncState &async_state);
-       void sync_corrector_finish(int lev, Parallel::AsyncSyncState &async_state, MyList<var> *VarList);
-       void sync_evolution(int lev, MyList<var> *VarList, Parallel::SyncCache *cache_array = 0);
-       void restrict_evolution(int lev, MyList<var> *src_var_list, MyList<var> *dst_var_list);
-       void outbdlow2hi_evolution(int lev, MyList<var> *src_var_list, MyList<var> *dst_var_list);
-
+       void testRestrict();
+       void testOutBd();
+       
+       bool check_Stdin_Abort(); 
+
        virtual void Setup_Initial_Data_Cao();
        virtual void Setup_Initial_Data_Lousto();
        virtual void Initialize();
-       virtual void Read_Ansorg();
-       virtual void Read_Pablo() {};
-       virtual void Compute_Psi4(int lev);
-       virtual void Step(int lev, int YN);
-       virtual void Interp_Constraint(bool infg);
-       virtual void Constraint_Out();
-       virtual void Compute_Constraint();
+       virtual void Read_Ansorg();
+       virtual void Read_Pablo() {};
+       void InvalidateSyncCaches();
+       virtual void Compute_Psi4(int lev);
+       virtual void Step(int lev, int YN);
+#ifdef USE_GPU
+       void Step_MainPath_GPU(int lev, int YN);
+#endif
+       virtual void Interp_Constraint(bool infg);
+       virtual void Constraint_Out();
+       virtual void Compute_Constraint();
 
 #ifdef With_AHF
 protected:

AMSS_NCKU_source/bssn_cuda_ops.h

@@ -0,0 +1,68 @@
+#ifndef BSSN_CUDA_OPS_H
+#define BSSN_CUDA_OPS_H
+
+int bssn_cuda_enforce_ga(int *ex,
+                         double *dxx, double *gxy, double *gxz,
+                         double *dyy, double *gyz, double *dzz,
+                         double *Axx, double *Axy, double *Axz,
+                         double *Ayy, double *Ayz, double *Azz);
+
+int bssn_cuda_rk4_boundary_var(int *ex, double dT,
+                               const double *X, const double *Y, const double *Z,
+                               double xmin, double ymin, double zmin,
+                               double xmax, double ymax, double zmax,
+                               const double *state0,
+                               const double *phi_field,
+                               const double *lap_field,
+                               const double *boundary_src,
+                               double *stage_data,
+                               double *rhs_accum,
+                               double propspeed,
+                               const double SoA[3],
+                               int symmetry,
+                               int lev,
+                               int rk_stage,
+                               bool force_host_boundary_fix,
+                               bool download_to_host = true);
+
+int bssn_cuda_rk4_boundary_batch(int *ex, double dT,
+                                 const double *X, const double *Y, const double *Z,
+                                 double xmin, double ymin, double zmin,
+                                 double xmax, double ymax, double zmax,
+                                 int symmetry,
+                                 const double *const *state0_list,
+                                 double *const *stage_data_list,
+                                 double *const *rhs_accum_list,
+                                 int num_var,
+                                 int rk_stage,
+                                 bool download_to_host = false);
+
+int bssn_cuda_lowerbound(int *ex, double *chi, double tinny, bool download_to_host = true);
+int bssn_cuda_download_buffer(int *ex, double *host_ptr);
+void bssn_cuda_release_rk4_caches();
+void bssn_cuda_release_interp_caches();
+
+int bssn_cuda_prolong3_pack(int wei,
+                            const double *llbc, const double *uubc, const int *extc, const double *func,
+                            const double *llbf, const double *uubf, const int *extf, double *funf,
+                            const double *llbp, const double *uubp,
+                            const double *SoA, int symmetry);
+
+int bssn_cuda_restrict3_pack(int wei,
+                             const double *llbc, const double *uubc, const int *extc, double *func,
+                             const double *llbf, const double *uubf, const int *extf, const double *funf,
+                             const double *llbr, const double *uubr,
+                             const double *SoA, int symmetry);
+
+int bssn_cuda_interp_points_batch(const int *ex,
+                                  const double *X, const double *Y, const double *Z,
+                                  const double *const *fields,
+                                  const double *soa_flat,
+                                  int num_var,
+                                  const double *px, const double *py, const double *pz,
+                                  int num_points,
+                                  int ordn,
+                                  int symmetry,
+                                  double *out);
+
+#endif

AMSS_NCKU_source/bssn_cuda_step.C

@@ -0,0 +1,936 @@
+#include "macrodef.h"
+
+#ifdef USE_GPU
+
+#include <algorithm>
+#include <cmath>
+#include <cstring>
+#include <cstdlib>
+#include <iomanip>
+#include <vector>
+
+#include "bssn_class.h"
+#include "bssn_cuda_ops.h"
+#include "bssn_gpu.h"
+#include "bssn_macro.h"
+
+namespace
+{
+enum StageProfileMetric
+{
+  STAGE_PROFILE_TOTAL = 0,
+  STAGE_PROFILE_RHS,
+  STAGE_PROFILE_RUN_STAGE,
+  STAGE_PROFILE_RUN_STAGE_DEVICE,
+  STAGE_PROFILE_RUN_STAGE_HOST_FIX,
+  STAGE_PROFILE_LOWERBOUND,
+  STAGE_PROFILE_ENSURE,
+  STAGE_PROFILE_DOWNLOAD,
+  STAGE_PROFILE_CLEAR_CACHE,
+  STAGE_PROFILE_SYNC_START,
+  STAGE_PROFILE_SYNC_FINISH,
+  STAGE_PROFILE_REFRESH,
+  STAGE_PROFILE_COUNT
+};
+
+static const int kStageProfileMaxLevels = 32;
+
+struct StageProfileStore
+{
+  bool env_checked;
+  bool enabled;
+  int calls[kStageProfileMaxLevels];
+  double metric[kStageProfileMaxLevels][STAGE_PROFILE_COUNT];
+};
+
+StageProfileStore &stage_profile_store()
+{
+  static StageProfileStore store = {};
+  return store;
+}
+
+bool stage_profile_enabled()
+{
+  StageProfileStore &store = stage_profile_store();
+  if (!store.env_checked)
+  {
+    const char *env = getenv("AMSS_GPU_STAGE_TIMING");
+    store.enabled = (env && env[0] && strcmp(env, "0") != 0);
+    store.env_checked = true;
+  }
+  return store.enabled;
+}
+
+void stage_profile_note_call(int lev)
+{
+  if (lev >= 0 && lev < kStageProfileMaxLevels)
+    stage_profile_store().calls[lev]++;
+}
+
+void stage_profile_add(int lev, StageProfileMetric metric, double seconds)
+{
+  if (lev >= 0 && lev < kStageProfileMaxLevels)
+    stage_profile_store().metric[lev][metric] += seconds;
+}
+
+const char *stage_profile_metric_name(StageProfileMetric metric)
+{
+  switch (metric)
+  {
+  case STAGE_PROFILE_TOTAL:
+    return "total";
+  case STAGE_PROFILE_RHS:
+    return "rhs";
+  case STAGE_PROFILE_RUN_STAGE:
+    return "run_stage";
+  case STAGE_PROFILE_RUN_STAGE_DEVICE:
+    return "run_stage_dev";
+  case STAGE_PROFILE_RUN_STAGE_HOST_FIX:
+    return "run_stage_host";
+  case STAGE_PROFILE_LOWERBOUND:
+    return "lower";
+  case STAGE_PROFILE_ENSURE:
+    return "ensure";
+  case STAGE_PROFILE_DOWNLOAD:
+    return "download";
+  case STAGE_PROFILE_CLEAR_CACHE:
+    return "clear_cache";
+  case STAGE_PROFILE_SYNC_START:
+    return "sync_start";
+  case STAGE_PROFILE_SYNC_FINISH:
+    return "sync_finish";
+  case STAGE_PROFILE_REFRESH:
+    return "refresh";
+  default:
+    return "unknown";
+  }
+}
+} // namespace
+
+void bssn_cuda_dump_stage_profile()
+{
+  if (!stage_profile_enabled())
+    return;
+
+  int myrank = 0;
+  MPI_Comm_rank(MPI_COMM_WORLD, &myrank);
+
+  StageProfileStore &store = stage_profile_store();
+  int global_calls_sum[kStageProfileMaxLevels] = {};
+  double global_metric_sum[kStageProfileMaxLevels][STAGE_PROFILE_COUNT] = {};
+  double global_metric_max[kStageProfileMaxLevels][STAGE_PROFILE_COUNT] = {};
+
+  MPI_Reduce(store.calls, global_calls_sum, kStageProfileMaxLevels, MPI_INT, MPI_SUM, 0, MPI_COMM_WORLD);
+  MPI_Reduce(store.metric[0], global_metric_sum[0],
+             kStageProfileMaxLevels * STAGE_PROFILE_COUNT,
+             MPI_DOUBLE, MPI_SUM, 0, MPI_COMM_WORLD);
+  MPI_Reduce(store.metric[0], global_metric_max[0],
+             kStageProfileMaxLevels * STAGE_PROFILE_COUNT,
+             MPI_DOUBLE, MPI_MAX, 0, MPI_COMM_WORLD);
+
+  if (myrank != 0)
+    return;
+
+  cout << endl;
+  cout << " GPU stage timing summary (sum/max over MPI ranks) " << endl;
+  cout << " lev  calls";
+  for (int metric = 0; metric < STAGE_PROFILE_COUNT; ++metric)
+    cout << "  " << setw(22) << stage_profile_metric_name(static_cast<StageProfileMetric>(metric));
+  cout << endl;
+
+  for (int lev = 0; lev < kStageProfileMaxLevels; ++lev)
+  {
+    if (global_calls_sum[lev] == 0)
+      continue;
+
+    cout << setw(4) << lev << "  " << setw(5) << global_calls_sum[lev];
+    for (int metric = 0; metric < STAGE_PROFILE_COUNT; ++metric)
+    {
+      cout << "  "
+           << setw(10) << setprecision(6) << fixed << global_metric_sum[lev][metric]
+           << "/"
+           << setw(10) << setprecision(6) << fixed << global_metric_max[lev][metric];
+    }
+    cout << endl;
+  }
+  cout << endl;
+}
+
+void bssn_class::Step_MainPath_GPU(int lev, int YN)
+{
+#ifdef WithShell
+#error "Step_MainPath_GPU currently supports Patch grids only."
+#endif
+
+  const bool profile_enabled = stage_profile_enabled();
+  const double step_total_begin = profile_enabled ? MPI_Wtime() : 0.0;
+  if (profile_enabled)
+    stage_profile_note_call(lev);
+
+  if (bssn_gpu_bind_process_device(myrank))
+  {
+    cerr << "GPU device bind failure on MPI rank " << myrank << endl;
+    MPI_Abort(MPI_COMM_WORLD, 1);
+  }
+  if (profile_enabled)
+  {
+    const double t0 = MPI_Wtime();
+    bssn_gpu_clear_cached_device_buffers();
+    stage_profile_add(lev, STAGE_PROFILE_CLEAR_CACHE, MPI_Wtime() - t0);
+  }
+  else
+    bssn_gpu_clear_cached_device_buffers();
+
+  setpbh(BH_num, Porg0, Mass, BH_num_input);
+
+  const double dT_lev = dT * pow(0.5, Mymax(lev, trfls));
+
+#if (MAPBH == 1)
+  if (BH_num > 0 && lev == GH->levels - 1)
+  {
+    compute_Porg_rhs(Porg0, Porg_rhs, Sfx0, Sfy0, Sfz0, lev);
+    for (int ithBH = 0; ithBH < BH_num; ithBH++)
+    {
+      for (int ith = 0; ith < 3; ith++)
+        Porg1[ithBH][ith] = Porg0[ithBH][ith] + Porg_rhs[ithBH][ith] * dT_lev;
+      if (Symmetry > 0)
+        Porg1[ithBH][2] = fabs(Porg1[ithBH][2]);
+      if (Symmetry == 2)
+      {
+        Porg1[ithBH][0] = fabs(Porg1[ithBH][0]);
+        Porg1[ithBH][1] = fabs(Porg1[ithBH][1]);
+      }
+    }
+  }
+
+  if (lev == a_lev)
+    AnalysisStuff(lev, dT_lev);
+#endif
+
+#ifdef With_AHF
+  AH_Step_Find(lev, dT_lev);
+#endif
+
+  const bool BB = fgt(PhysTime, StartTime, dT_lev / 2);
+  (void)BB;
+  double ndeps = (lev < GH->movls) ? numepsb : numepss;
+  double TRK4 = PhysTime;
+  int iter_count = 0;
+  int pre = 0, cor = 1;
+  int ERROR = 0;
+  const bool keep_stage_sync_on_device = (RPS == 1) && (MAPBH == 1) && (REGLEV == 0);
+
+  auto run_stage_on_block =
+      [&](Block *cg, Patch *patch, MyList<var> *state0_list,
+          MyList<var> *boundary_src_list, MyList<var> *stage_data_list,
+          MyList<var> *rhs_list, int rk_stage) {
+        MyList<var> *varl0 = state0_list;
+        MyList<var> *varlb = boundary_src_list;
+        MyList<var> *varls = stage_data_list;
+        MyList<var> *varlr = rhs_list;
+        std::vector<const double *> batch_state0;
+        std::vector<double *> batch_stage;
+        std::vector<double *> batch_rhs;
+
+        while (varl0)
+        {
+          const bool force_host_boundary_fix = false;
+          const bool can_batch_device_path = (lev > 0) && !force_host_boundary_fix;
+          if (can_batch_device_path)
+          {
+            batch_state0.push_back(cg->fgfs[varl0->data->sgfn]);
+            batch_stage.push_back(cg->fgfs[varls->data->sgfn]);
+            batch_rhs.push_back(cg->fgfs[varlr->data->sgfn]);
+            varl0 = varl0->next;
+            varlb = varlb->next;
+            varls = varls->next;
+            varlr = varlr->next;
+            continue;
+          }
+
+          const double var_begin = profile_enabled ? MPI_Wtime() : 0.0;
+          if (bssn_cuda_rk4_boundary_var(cg->shape, dT_lev,
+                                         cg->X[0], cg->X[1], cg->X[2],
+                                         patch->bbox[0], patch->bbox[1], patch->bbox[2],
+                                         patch->bbox[3], patch->bbox[4], patch->bbox[5],
+                                         cg->fgfs[varl0->data->sgfn],
+                                         cg->fgfs[phi0->sgfn],
+                                         cg->fgfs[Lap0->sgfn],
+                                         cg->fgfs[varlb->data->sgfn],
+                                         cg->fgfs[varls->data->sgfn],
+                                         cg->fgfs[varlr->data->sgfn],
+                                         varl0->data->propspeed,
+                                         varl0->data->SoA,
+                                         Symmetry, lev, rk_stage,
+                                         force_host_boundary_fix, false))
+          {
+            cerr << "GPU rk4/boundary failure: lev=" << lev
+                 << " rk_stage=" << rk_stage
+                 << " var=" << varl0->data->name
+                 << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                 << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                 << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+            ERROR = 1;
+            break;
+          }
+          if (profile_enabled)
+          {
+            stage_profile_add(lev,
+                              force_host_boundary_fix ? STAGE_PROFILE_RUN_STAGE_HOST_FIX
+                                                      : STAGE_PROFILE_RUN_STAGE_DEVICE,
+                              MPI_Wtime() - var_begin);
+          }
+          varl0 = varl0->next;
+          varlb = varlb->next;
+          varls = varls->next;
+          varlr = varlr->next;
+        }
+
+        if (!ERROR && !batch_state0.empty())
+        {
+          const double batch_begin = profile_enabled ? MPI_Wtime() : 0.0;
+          if (bssn_cuda_rk4_boundary_batch(cg->shape, dT_lev,
+                                           cg->X[0], cg->X[1], cg->X[2],
+                                           patch->bbox[0], patch->bbox[1], patch->bbox[2],
+                                           patch->bbox[3], patch->bbox[4], patch->bbox[5],
+                                           Symmetry,
+                                           &batch_state0[0],
+                                           &batch_stage[0],
+                                           &batch_rhs[0],
+                                           static_cast<int>(batch_state0.size()),
+                                           rk_stage, false))
+          {
+            cerr << "GPU rk4/boundary batch failure: lev=" << lev
+                 << " rk_stage=" << rk_stage
+                 << " vars=" << batch_state0.size()
+                 << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                 << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                 << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+            ERROR = 1;
+          }
+          else if (profile_enabled)
+          {
+            stage_profile_add(lev, STAGE_PROFILE_RUN_STAGE_DEVICE, MPI_Wtime() - batch_begin);
+          }
+        }
+      };
+
+  auto stage_download_var_list =
+      [&](Block *cg, MyList<var> *var_list, bool skip_unmapped) {
+        std::vector<double *> batch_host_ptrs;
+        std::vector<MyList<var> *> batch_vars;
+        while (var_list)
+        {
+          double *host_ptr = cg->fgfs[var_list->data->sgfn];
+          if (skip_unmapped && !bssn_gpu_find_device_buffer(host_ptr))
+          {
+            var_list = var_list->next;
+            continue;
+          }
+          batch_host_ptrs.push_back(host_ptr);
+          batch_vars.push_back(var_list);
+          var_list = var_list->next;
+        }
+        if (!batch_host_ptrs.empty() &&
+            bssn_gpu_download_buffer_batch(cg->shape, &batch_host_ptrs[0],
+                                           static_cast<int>(batch_host_ptrs.size())))
+        {
+          for (size_t i = 0; i < batch_host_ptrs.size(); ++i)
+          {
+            if (bssn_cuda_download_buffer(cg->shape, batch_host_ptrs[i]))
+            {
+              cerr << "GPU stage download failure: lev=" << lev
+                   << " var=" << batch_vars[i]->data->name
+                   << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                   << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                   << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+              ERROR = 1;
+              break;
+            }
+          }
+        }
+      };
+
+  auto stage_download_patch_list =
+      [&](MyList<var> *var_list, bool skip_unmapped) {
+        MyList<Patch> *patch_it = GH->PatL[lev];
+        while (patch_it)
+        {
+          MyList<Block> *block_it = patch_it->data->blb;
+          while (block_it)
+          {
+            Block *cg = block_it->data;
+            if (myrank == cg->rank)
+              stage_download_var_list(cg, var_list, skip_unmapped);
+
+            if (block_it == patch_it->data->ble)
+              break;
+            block_it = block_it->next;
+          }
+          if (ERROR)
+            break;
+          patch_it = patch_it->next;
+        }
+      };
+
+  auto ensure_stage_device_var_list =
+      [&](Block *cg, MyList<var> *var_list) {
+        const int n = cg->shape[0] * cg->shape[1] * cg->shape[2];
+        while (var_list)
+        {
+          double *host_ptr = cg->fgfs[var_list->data->sgfn];
+          if (!bssn_gpu_find_device_buffer(host_ptr) &&
+              bssn_gpu_stage_upload_buffer(host_ptr, n))
+          {
+            cerr << "GPU state ensure failure: lev=" << lev
+                 << " var=" << var_list->data->name
+                 << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                 << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                 << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+            ERROR = 1;
+            break;
+          }
+          var_list = var_list->next;
+        }
+      };
+
+  auto refresh_synced_device_regions =
+      [&](Block *cg, MyList<var> *var_list, Parallel::SyncCache &cache) {
+        std::vector<Parallel::gridseg *> local_segments;
+        for (int node = 0; node < cache.cpusize; ++node)
+        {
+          MyList<Parallel::gridseg> *seg = cache.combined_dst[node];
+          while (seg)
+          {
+            if (seg->data && seg->data->Bg == cg)
+              local_segments.push_back(seg->data);
+            seg = seg->next;
+          }
+        }
+
+        if (local_segments.empty())
+          return;
+
+        const int n = cg->shape[0] * cg->shape[1] * cg->shape[2];
+        while (var_list)
+        {
+          double *host_ptr = cg->fgfs[var_list->data->sgfn];
+          if (!bssn_gpu_find_device_buffer(host_ptr))
+          {
+            if (bssn_gpu_stage_upload_buffer(host_ptr, n))
+            {
+              cerr << "GPU sync refresh upload failure: lev=" << lev
+                   << " var=" << var_list->data->name
+                   << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                   << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                   << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+              ERROR = 1;
+              break;
+            }
+          }
+          else
+          {
+            for (size_t i = 0; i < local_segments.size(); ++i)
+            {
+              Parallel::gridseg *seg = local_segments[i];
+              if (bssn_gpu_stage_upload_region(host_ptr,
+                                               cg->shape,
+                                               cg->bbox,
+                                               cg->bbox + dim,
+                                               seg->shape,
+                                               seg->llb))
+              {
+                cerr << "GPU sync region refresh failure: lev=" << lev
+                     << " var=" << var_list->data->name
+                     << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                     << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                     << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+                ERROR = 1;
+                break;
+              }
+            }
+            if (ERROR)
+              break;
+          }
+          var_list = var_list->next;
+        }
+      };
+
+  auto refresh_stage_device_after_sync =
+      [&](MyList<var> *var_list, Parallel::SyncCache &cache) {
+        MyList<Patch> *patch_it = GH->PatL[lev];
+        while (patch_it)
+        {
+          MyList<Block> *block_it = patch_it->data->blb;
+          while (block_it)
+          {
+            Block *cg = block_it->data;
+            if (myrank == cg->rank)
+              refresh_synced_device_regions(cg, var_list, cache);
+
+            if (block_it == patch_it->data->ble)
+              break;
+            block_it = block_it->next;
+          }
+          if (ERROR)
+            break;
+          patch_it = patch_it->next;
+        }
+      };
+
+  auto refresh_stage_host_before_sync =
+      [&](MyList<var> *var_list, Parallel::SyncCache &cache) -> bool {
+        if (!cache.valid || !cache.combined_src || myrank < 0 || myrank >= cache.cpusize)
+          return false;
+
+        MyList<Patch> *patch_it = GH->PatL[lev];
+        while (patch_it)
+        {
+          MyList<Block> *block_it = patch_it->data->blb;
+          while (block_it)
+          {
+            Block *cg = block_it->data;
+            if (myrank == cg->rank)
+            {
+              std::vector<Parallel::gridseg *> local_segments;
+              MyList<Parallel::gridseg> *seg = cache.combined_src[myrank];
+              while (seg)
+              {
+                if (seg->data && seg->data->Bg == cg)
+                  local_segments.push_back(seg->data);
+                seg = seg->next;
+              }
+
+              if (!local_segments.empty())
+              {
+                MyList<var> *var_it = var_list;
+                while (var_it)
+                {
+                  double *host_ptr = cg->fgfs[var_it->data->sgfn];
+                  for (size_t i = 0; i < local_segments.size(); ++i)
+                  {
+                    Parallel::gridseg *src_seg = local_segments[i];
+                    if (bssn_gpu_stage_download_region(host_ptr,
+                                                       cg->shape,
+                                                       cg->bbox,
+                                                       cg->bbox + dim,
+                                                       src_seg->shape,
+                                                       src_seg->llb))
+                    {
+                      cerr << "GPU sync region download failure: lev=" << lev
+                           << " var=" << var_it->data->name
+                           << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                           << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                           << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+                      ERROR = 1;
+                      return true;
+                    }
+                  }
+                  var_it = var_it->next;
+                }
+              }
+            }
+
+            if (block_it == patch_it->data->ble)
+              break;
+            block_it = block_it->next;
+          }
+          patch_it = patch_it->next;
+        }
+
+        return true;
+      };
+
+  auto can_pack_sync_from_device =
+      [&](MyList<var> *var_list, Parallel::SyncCache &cache) -> bool {
+        if (!cache.valid || !cache.combined_src || myrank < 0 || myrank >= cache.cpusize)
+          return false;
+
+        MyList<Parallel::gridseg> *seg = cache.combined_src[myrank];
+        while (seg)
+        {
+          MyList<var> *var_it = var_list;
+          while (var_it)
+          {
+            if (!bssn_gpu_find_device_buffer(seg->data->Bg->fgfs[var_it->data->sgfn]))
+              return false;
+            var_it = var_it->next;
+          }
+          seg = seg->next;
+        }
+        return true;
+      };
+
+  MyList<Patch> *Pp = GH->PatL[lev];
+  while (Pp)
+  {
+    MyList<Block> *BP = Pp->data->blb;
+    while (BP)
+    {
+      Block *cg = BP->data;
+      if (myrank == cg->rank)
+      {
+        double t0 = 0.0;
+        if (profile_enabled)
+          t0 = MPI_Wtime();
+        if (gpu_rhs(CALLED_BY_STEP, myrank, RHS_PARA_CALLED_FIRST_TIME))
+          ERROR = 1;
+        if (profile_enabled)
+          stage_profile_add(lev, STAGE_PROFILE_RHS, MPI_Wtime() - t0);
+
+        if (profile_enabled)
+          t0 = MPI_Wtime();
+        run_stage_on_block(cg, Pp->data, StateList, StateList, SynchList_pre, RHSList, iter_count);
+        if (profile_enabled)
+          stage_profile_add(lev, STAGE_PROFILE_RUN_STAGE, MPI_Wtime() - t0);
+
+        if (profile_enabled)
+          t0 = MPI_Wtime();
+        if (bssn_cuda_lowerbound(cg->shape, cg->fgfs[phi->sgfn], chitiny, false))
+        {
+          cerr << "GPU lowerbound failure: lev=" << lev
+               << " rk_stage=" << iter_count
+               << " var=" << phi->name
+               << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+               << cg->bbox[1] << ":" << cg->bbox[4] << ","
+               << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+          ERROR = 1;
+        }
+        if (profile_enabled)
+          stage_profile_add(lev, STAGE_PROFILE_LOWERBOUND, MPI_Wtime() - t0);
+      }
+      if (BP == Pp->data->ble)
+        break;
+      BP = BP->next;
+    }
+    Pp = Pp->next;
+  }
+
+  if (!ERROR)
+  {
+    if (!keep_stage_sync_on_device)
+    {
+      double t0 = 0.0;
+      if (profile_enabled)
+        t0 = MPI_Wtime();
+      stage_download_patch_list(SynchList_pre, false);
+      if (profile_enabled)
+        stage_profile_add(lev, STAGE_PROFILE_DOWNLOAD, MPI_Wtime() - t0);
+      if (!ERROR)
+      {
+        if (profile_enabled)
+          t0 = MPI_Wtime();
+        bssn_gpu_clear_cached_device_buffers();
+        if (profile_enabled)
+          stage_profile_add(lev, STAGE_PROFILE_CLEAR_CACHE, MPI_Wtime() - t0);
+      }
+    }
+  }
+
+  MPI_Request err_req_pre;
+  {
+    int erh = ERROR;
+    MPI_Iallreduce(&erh, &ERROR, 1, MPI_INT, MPI_SUM, MPI_COMM_WORLD, &err_req_pre);
+  }
+
+  Parallel::AsyncSyncState async_pre;
+  if (profile_enabled)
+  {
+    const double t0 = MPI_Wtime();
+    Parallel::Sync_start(GH->PatL[lev], SynchList_pre, Symmetry, sync_cache_pre[lev], async_pre);
+    stage_profile_add(lev, STAGE_PROFILE_SYNC_START, MPI_Wtime() - t0);
+  }
+  else
+    Parallel::Sync_start(GH->PatL[lev], SynchList_pre, Symmetry, sync_cache_pre[lev], async_pre);
+  if (profile_enabled)
+  {
+    const double t0 = MPI_Wtime();
+    Parallel::Sync_finish(sync_cache_pre[lev], async_pre, SynchList_pre, Symmetry,
+                          !keep_stage_sync_on_device);
+    stage_profile_add(lev, STAGE_PROFILE_SYNC_FINISH, MPI_Wtime() - t0);
+  }
+  else
+    Parallel::Sync_finish(sync_cache_pre[lev], async_pre, SynchList_pre, Symmetry,
+                          !keep_stage_sync_on_device);
+  if (!ERROR && !keep_stage_sync_on_device)
+  {
+    if (profile_enabled)
+    {
+      const double t0 = MPI_Wtime();
+      refresh_stage_device_after_sync(SynchList_pre, sync_cache_pre[lev]);
+      stage_profile_add(lev, STAGE_PROFILE_REFRESH, MPI_Wtime() - t0);
+    }
+    else
+      refresh_stage_device_after_sync(SynchList_pre, sync_cache_pre[lev]);
+  }
+
+  MPI_Wait(&err_req_pre, MPI_STATUS_IGNORE);
+  if (ERROR)
+  {
+    Parallel::Dump_Data(GH->PatL[lev], StateList, 0, PhysTime, dT_lev);
+    if (myrank == 0)
+    {
+      if (ErrorMonitor->outfile)
+        ErrorMonitor->outfile << "find NaN in state variables at t = " << PhysTime
+                              << ", lev = " << lev << endl;
+      MPI_Abort(MPI_COMM_WORLD, 1);
+    }
+  }
+
+#if (MAPBH == 0)
+  if (BH_num > 0 && lev == GH->levels - 1)
+  {
+    compute_Porg_rhs(Porg0, Porg_rhs, Sfx0, Sfy0, Sfz0, lev);
+    for (int ithBH = 0; ithBH < BH_num; ithBH++)
+    {
+      f_rungekutta4_scalar(dT_lev, Porg0[ithBH][0], Porg[ithBH][0], Porg_rhs[ithBH][0], iter_count);
+      f_rungekutta4_scalar(dT_lev, Porg0[ithBH][1], Porg[ithBH][1], Porg_rhs[ithBH][1], iter_count);
+      f_rungekutta4_scalar(dT_lev, Porg0[ithBH][2], Porg[ithBH][2], Porg_rhs[ithBH][2], iter_count);
+      if (Symmetry > 0)
+        Porg[ithBH][2] = fabs(Porg[ithBH][2]);
+      if (Symmetry == 2)
+      {
+        Porg[ithBH][0] = fabs(Porg[ithBH][0]);
+        Porg[ithBH][1] = fabs(Porg[ithBH][1]);
+      }
+    }
+  }
+
+  if (lev == a_lev)
+    AnalysisStuff(lev, dT_lev);
+#endif
+
+  for (iter_count = 1; iter_count < 4; iter_count++)
+  {
+    if (iter_count == 1 || iter_count == 3)
+      TRK4 += dT_lev / 2;
+
+    Pp = GH->PatL[lev];
+    while (Pp)
+    {
+      MyList<Block> *BP = Pp->data->blb;
+      while (BP)
+      {
+        Block *cg = BP->data;
+        if (myrank == cg->rank)
+        {
+          double t0 = 0.0;
+          if (profile_enabled)
+            t0 = MPI_Wtime();
+          ensure_stage_device_var_list(cg, SynchList_pre);
+          if (profile_enabled)
+            stage_profile_add(lev, STAGE_PROFILE_ENSURE, MPI_Wtime() - t0);
+
+          if (profile_enabled)
+            t0 = MPI_Wtime();
+          if (gpu_rhs(CALLED_BY_STEP, myrank, RHS_PARA_CALLED_THEN))
+            ERROR = 1;
+          if (profile_enabled)
+            stage_profile_add(lev, STAGE_PROFILE_RHS, MPI_Wtime() - t0);
+
+          if (profile_enabled)
+            t0 = MPI_Wtime();
+          run_stage_on_block(cg, Pp->data, StateList, SynchList_pre, SynchList_cor, RHSList, iter_count);
+          if (profile_enabled)
+            stage_profile_add(lev, STAGE_PROFILE_RUN_STAGE, MPI_Wtime() - t0);
+
+          if (profile_enabled)
+            t0 = MPI_Wtime();
+          if (bssn_cuda_lowerbound(cg->shape, cg->fgfs[phi1->sgfn], chitiny, false))
+          {
+          cerr << "GPU lowerbound failure: lev=" << lev
+               << " rk_stage=" << iter_count
+               << " var=" << phi1->name
+                 << " bbox=(" << cg->bbox[0] << ":" << cg->bbox[3] << ","
+                 << cg->bbox[1] << ":" << cg->bbox[4] << ","
+                 << cg->bbox[2] << ":" << cg->bbox[5] << ")" << endl;
+            ERROR = 1;
+          }
+          if (profile_enabled)
+            stage_profile_add(lev, STAGE_PROFILE_LOWERBOUND, MPI_Wtime() - t0);
+        }
+
+        if (BP == Pp->data->ble)
+          break;
+        BP = BP->next;
+      }
+      Pp = Pp->next;
+    }
+
+    if (!ERROR)
+    {
+      if (!keep_stage_sync_on_device)
+      {
+        double t0 = 0.0;
+        if (profile_enabled)
+          t0 = MPI_Wtime();
+        stage_download_patch_list(SynchList_cor, false);
+        if (profile_enabled)
+          stage_profile_add(lev, STAGE_PROFILE_DOWNLOAD, MPI_Wtime() - t0);
+        if (!ERROR)
+        {
+          if (profile_enabled)
+            t0 = MPI_Wtime();
+          bssn_gpu_clear_cached_device_buffers();
+          if (profile_enabled)
+            stage_profile_add(lev, STAGE_PROFILE_CLEAR_CACHE, MPI_Wtime() - t0);
+        }
+      }
+    }
+
+    MPI_Request err_req_cor;
+    {
+      int erh = ERROR;
+      MPI_Iallreduce(&erh, &ERROR, 1, MPI_INT, MPI_SUM, MPI_COMM_WORLD, &err_req_cor);
+    }
+
+    Parallel::AsyncSyncState async_cor;
+    if (profile_enabled)
+    {
+      const double t0 = MPI_Wtime();
+      Parallel::Sync_start(GH->PatL[lev], SynchList_cor, Symmetry, sync_cache_cor[lev], async_cor);
+      stage_profile_add(lev, STAGE_PROFILE_SYNC_START, MPI_Wtime() - t0);
+    }
+    else
+      Parallel::Sync_start(GH->PatL[lev], SynchList_cor, Symmetry, sync_cache_cor[lev], async_cor);
+    if (profile_enabled)
+    {
+      const double t0 = MPI_Wtime();
+      Parallel::Sync_finish(sync_cache_cor[lev], async_cor, SynchList_cor, Symmetry,
+                            !keep_stage_sync_on_device);
+      stage_profile_add(lev, STAGE_PROFILE_SYNC_FINISH, MPI_Wtime() - t0);
+    }
+    else
+      Parallel::Sync_finish(sync_cache_cor[lev], async_cor, SynchList_cor, Symmetry,
+                            !keep_stage_sync_on_device);
+    if (!ERROR && !keep_stage_sync_on_device && iter_count < 3)
+    {
+      if (profile_enabled)
+      {
+        const double t0 = MPI_Wtime();
+        refresh_stage_device_after_sync(SynchList_cor, sync_cache_cor[lev]);
+        stage_profile_add(lev, STAGE_PROFILE_REFRESH, MPI_Wtime() - t0);
+      }
+      else
+        refresh_stage_device_after_sync(SynchList_cor, sync_cache_cor[lev]);
+    }
+
+    MPI_Wait(&err_req_cor, MPI_STATUS_IGNORE);
+    if (ERROR)
+    {
+      Parallel::Dump_Data(GH->PatL[lev], SynchList_pre, 0, PhysTime, dT_lev);
+      if (myrank == 0)
+      {
+        if (ErrorMonitor->outfile)
+          ErrorMonitor->outfile << "find NaN in RK4 substep#" << iter_count
+                                << " variables at t = " << PhysTime
+                                << ", lev = " << lev << endl;
+        MPI_Abort(MPI_COMM_WORLD, 1);
+      }
+    }
+
+#if (MAPBH == 0)
+    if (BH_num > 0 && lev == GH->levels - 1)
+    {
+      compute_Porg_rhs(Porg, Porg1, Sfx, Sfy, Sfz, lev);
+      for (int ithBH = 0; ithBH < BH_num; ithBH++)
+      {
+        f_rungekutta4_scalar(dT_lev, Porg0[ithBH][0], Porg1[ithBH][0], Porg_rhs[ithBH][0], iter_count);
+        f_rungekutta4_scalar(dT_lev, Porg0[ithBH][1], Porg1[ithBH][1], Porg_rhs[ithBH][1], iter_count);
+        f_rungekutta4_scalar(dT_lev, Porg0[ithBH][2], Porg1[ithBH][2], Porg_rhs[ithBH][2], iter_count);
+        if (Symmetry > 0)
+          Porg1[ithBH][2] = fabs(Porg1[ithBH][2]);
+        if (Symmetry == 2)
+        {
+          Porg1[ithBH][0] = fabs(Porg1[ithBH][0]);
+          Porg1[ithBH][1] = fabs(Porg1[ithBH][1]);
+        }
+      }
+    }
+#endif
+
+    if (iter_count < 3)
+    {
+      Pp = GH->PatL[lev];
+      while (Pp)
+      {
+        MyList<Block> *BP = Pp->data->blb;
+        while (BP)
+        {
+          BP->data->swapList(SynchList_pre, SynchList_cor, myrank);
+          if (BP == Pp->data->ble)
+            break;
+          BP = BP->next;
+        }
+        Pp = Pp->next;
+      }
+
+#if (MAPBH == 0)
+      if (BH_num > 0 && lev == GH->levels - 1)
+      {
+        for (int ithBH = 0; ithBH < BH_num; ithBH++)
+        {
+          Porg[ithBH][0] = Porg1[ithBH][0];
+          Porg[ithBH][1] = Porg1[ithBH][1];
+          Porg[ithBH][2] = Porg1[ithBH][2];
+        }
+      }
+#endif
+    }
+  }
+
+#if (RPS == 0)
+  RestrictProlong(lev, YN, BB);
+#endif
+
+  Pp = GH->PatL[lev];
+  while (Pp)
+  {
+    MyList<Block> *BP = Pp->data->blb;
+    while (BP)
+    {
+      Block *cg = BP->data;
+      cg->swapList(StateList, SynchList_cor, myrank);
+      cg->swapList(OldStateList, SynchList_cor, myrank);
+      if (BP == Pp->data->ble)
+        break;
+      BP = BP->next;
+    }
+    Pp = Pp->next;
+  }
+
+  if (!ERROR && keep_stage_sync_on_device)
+  {
+    // After the swaps above, only StateList points at arrays updated during this step.
+    // OldStateList/SynchList_cor remain valid on host because their backing arrays were
+    // read-only during the RK step, and SynchList_pre is reused only as scratch later.
+    const double t0 = profile_enabled ? MPI_Wtime() : 0.0;
+    stage_download_patch_list(StateList, true);
+    if (profile_enabled)
+      stage_profile_add(lev, STAGE_PROFILE_DOWNLOAD, MPI_Wtime() - t0);
+  }
+
+  if (profile_enabled)
+  {
+    const double t0 = MPI_Wtime();
+    bssn_gpu_clear_cached_device_buffers();
+    stage_profile_add(lev, STAGE_PROFILE_CLEAR_CACHE, MPI_Wtime() - t0);
+  }
+  else
+    bssn_gpu_clear_cached_device_buffers();
+
+  if (BH_num > 0 && lev == GH->levels - 1)
+  {
+    for (int ithBH = 0; ithBH < BH_num; ithBH++)
+    {
+      Porg0[ithBH][0] = Porg1[ithBH][0];
+      Porg0[ithBH][1] = Porg1[ithBH][1];
+      Porg0[ithBH][2] = Porg1[ithBH][2];
+    }
+  }
+
+  if (profile_enabled)
+    stage_profile_add(lev, STAGE_PROFILE_TOTAL, MPI_Wtime() - step_total_begin);
+}
+
+#endif

AMSS_NCKU_source/bssn_em_rhs_c.C

@@ -1,323 +0,0 @@
-#include "macrodef.h"
-#include "bssn_rhs.h"
-#include "share_func.h"
-#include "tool.h"
-#include <cstddef>
-
-/*
- * C 版 BSSN-EM RHS kernel — replaces empart.f90 + bssn_rhs.f90 for BSSN+Maxwell.
- *
- * Computes:
- *   1. All metric and EM field derivatives
- *   2. Physical metric, Christoffel-like terms
- *   3. EM field RHS (E, B, Kpsi, Kphi)
- *   4. Stress-energy tensor (rho, Si, Sij)
- *   5. Calls f_compute_rhs_bssn (C BSSN RHS) with stress-energy
- *   6. Advection + KO dissipation for EM fields
- *   7. NaN check
- */
-int f_compute_rhs_bssn_em_c(int *ex, double &T,
-                            double *X, double *Y, double *Z,
-                            double *chi, double *trK,
-                            double *dxx, double *gxy, double *gxz, double *dyy, double *gyz, double *dzz,
-                            double *Axx, double *Axy, double *Axz, double *Ayy, double *Ayz, double *Azz,
-                            double *Gamx, double *Gamy, double *Gamz,
-                            double *Lap, double *betax, double *betay, double *betaz,
-                            double *dtSfx, double *dtSfy, double *dtSfz,
-                            double *Ex,  double *Ey,  double *Ez,
-                            double *Bx,  double *By,  double *Bz,
-                            double *Kpsi, double *Kphi,
-                            double *Jx, double *Jy, double *Jz, double *qchar,
-                            double *chi_rhs, double *trK_rhs,
-                            double *gxx_rhs, double *gxy_rhs, double *gxz_rhs,
-                            double *gyy_rhs, double *gyz_rhs, double *gzz_rhs,
-                            double *Axx_rhs, double *Axy_rhs, double *Axz_rhs,
-                            double *Ayy_rhs, double *Ayz_rhs, double *Azz_rhs,
-                            double *Gamx_rhs, double *Gamy_rhs, double *Gamz_rhs,
-                            double *Lap_rhs, double *betax_rhs, double *betay_rhs, double *betaz_rhs,
-                            double *dtSfx_rhs, double *dtSfy_rhs, double *dtSfz_rhs,
-                            double *Ex_rhs,  double *Ey_rhs,  double *Ez_rhs,
-                            double *Bx_rhs,  double *By_rhs,  double *Bz_rhs,
-                            double *Kpsi_rhs, double *Kphi_rhs,
-                            double *rho, double *Sx, double *Sy, double *Sz,
-                            double *Sxx, double *Sxy, double *Sxz,
-                            double *Syy, double *Syz, double *Szz,
-                            double *Gamxxx, double *Gamxxy, double *Gamxxz,
-                            double *Gamxyy, double *Gamxyz, double *Gamxzz,
-                            double *Gamyxx, double *Gamyxy, double *Gamyxz,
-                            double *Gamyyy, double *Gamyyz, double *Gamyzz,
-                            double *Gamzxx, double *Gamzxy, double *Gamzxz,
-                            double *Gamzyy, double *Gamzyz, double *Gamzzz,
-                            double *Rxx, double *Rxy, double *Rxz,
-                            double *Ryy, double *Ryz, double *Rzz,
-                            double *ham_Res, double *movx_Res, double *movy_Res, double *movz_Res,
-                            double *Gmx_Res, double *Gmy_Res, double *Gmz_Res,
-                            int &Symmetry, int &Lev, double &eps, int &co)
-{
-    (void)T;
-    int gont = 0;
-    const int nx = ex[0], ny = ex[1], nz = ex[2];
-    const int all = nx * ny * nz;
-    const size_t n = (size_t)all;
-
-    const double ZEO = 0.0, ONE = 1.0, TWO = 2.0, FOUR = 4.0, EIT = 8.0;
-    const double HALF = 0.5, THR = 3.0, F3o2 = 1.5, PI = 3.14159265358979323846;
-    const double SYM = 1.0, ANTI = -1.0;
-    const double kappa = 1.0;
-    const double SSS[3]={SYM,SYM,SYM},   AAS[3]={ANTI,ANTI,SYM};
-    const double ASA[3]={ANTI,SYM,ANTI}, SAA[3]={SYM,ANTI,ANTI};
-    const double ASS[3]={ANTI,SYM,SYM},   SAS[3]={SYM,ANTI,SYM};
-    const double SSA[3]={SYM,SYM,ANTI};
-
-    /* ---- allocate temporary arrays ---- */
-    double *chix = (double*)malloc(n*sizeof(double));
-    double *chiy = (double*)malloc(n*sizeof(double));
-    double *chiz = (double*)malloc(n*sizeof(double));
-    double *Exx=(double*)malloc(n*sizeof(double)),*Exy=(double*)malloc(n*sizeof(double)),*Exz=(double*)malloc(n*sizeof(double));
-    double *Eyx=(double*)malloc(n*sizeof(double)),*Eyy=(double*)malloc(n*sizeof(double)),*Eyz=(double*)malloc(n*sizeof(double));
-    double *Ezx=(double*)malloc(n*sizeof(double)),*Ezy=(double*)malloc(n*sizeof(double)),*Ezz=(double*)malloc(n*sizeof(double));
-    double *Bxx=(double*)malloc(n*sizeof(double)),*Bxy=(double*)malloc(n*sizeof(double)),*Bxz=(double*)malloc(n*sizeof(double));
-    double *Byx=(double*)malloc(n*sizeof(double)),*Byy=(double*)malloc(n*sizeof(double)),*Byz=(double*)malloc(n*sizeof(double));
-    double *Bzx=(double*)malloc(n*sizeof(double)),*Bzy=(double*)malloc(n*sizeof(double)),*Bzz=(double*)malloc(n*sizeof(double));
-    double *Kpsix=(double*)malloc(n*sizeof(double)),*Kpsiy=(double*)malloc(n*sizeof(double)),*Kpsiz=(double*)malloc(n*sizeof(double));
-    double *Kphix=(double*)malloc(n*sizeof(double)),*Kphiy=(double*)malloc(n*sizeof(double)),*Kphiz=(double*)malloc(n*sizeof(double));
-    double *Lapx=(double*)malloc(n*sizeof(double)),*Lapy=(double*)malloc(n*sizeof(double)),*Lapz=(double*)malloc(n*sizeof(double));
-    double *betaxx=(double*)malloc(n*sizeof(double)),*betaxy=(double*)malloc(n*sizeof(double)),*betaxz=(double*)malloc(n*sizeof(double));
-    double *betayx=(double*)malloc(n*sizeof(double)),*betayy=(double*)malloc(n*sizeof(double)),*betayz=(double*)malloc(n*sizeof(double));
-    double *betazx=(double*)malloc(n*sizeof(double)),*betazy=(double*)malloc(n*sizeof(double)),*betazz=(double*)malloc(n*sizeof(double));
-    double *gxxx=(double*)malloc(n*sizeof(double)),*gxxy=(double*)malloc(n*sizeof(double)),*gxxz=(double*)malloc(n*sizeof(double));
-    double *gxyx=(double*)malloc(n*sizeof(double)),*gxyy=(double*)malloc(n*sizeof(double)),*gxyz=(double*)malloc(n*sizeof(double));
-    double *gxzx=(double*)malloc(n*sizeof(double)),*gxzy=(double*)malloc(n*sizeof(double)),*gxzz=(double*)malloc(n*sizeof(double));
-    double *gyyx=(double*)malloc(n*sizeof(double)),*gyyy=(double*)malloc(n*sizeof(double)),*gyyz=(double*)malloc(n*sizeof(double));
-    double *gyzx=(double*)malloc(n*sizeof(double)),*gyzy=(double*)malloc(n*sizeof(double)),*gyzz=(double*)malloc(n*sizeof(double));
-    double *gzzx=(double*)malloc(n*sizeof(double)),*gzzy=(double*)malloc(n*sizeof(double)),*gzzz=(double*)malloc(n*sizeof(double));
-    double *gupxx=(double*)malloc(n*sizeof(double)),*gupxy=(double*)malloc(n*sizeof(double)),*gupxz=(double*)malloc(n*sizeof(double));
-    double *gupyy=(double*)malloc(n*sizeof(double)),*gupyz=(double*)malloc(n*sizeof(double)),*gupzz=(double*)malloc(n*sizeof(double));
-
-    if (!chix||!chiy||!chiz||!Exx||!Exy||!Exz||!Eyx||!Eyy||!Eyz||!Ezx||!Ezy||!Ezz||
-        !Bxx||!Bxy||!Bxz||!Byx||!Byy||!Byz||!Bzx||!Bzy||!Bzz||
-        !Kpsix||!Kpsiy||!Kpsiz||!Kphix||!Kphiy||!Kphiz||
-        !Lapx||!Lapy||!Lapz||
-        !betaxx||!betaxy||!betaxz||!betayx||!betayy||!betayz||!betazx||!betazy||!betazz||
-        !gxxx||!gxxy||!gxxz||!gxyx||!gxyy||!gxyz||!gxzx||!gxzy||!gxzz||
-        !gyyx||!gyyy||!gyyz||!gyzx||!gyzy||!gyzz||!gzzx||!gzzy||!gzzz||
-        !gupxx||!gupxy||!gupxz||!gupyy||!gupyz||!gupzz) {
-        gont = 1;
-    }
-
-    /* ==== 1. Compute all derivatives ==== */
-    if (!gont) {
-
-    /* metric derivatives */
-    fderivs(ex, Lap, Lapx, Lapy, Lapz, X, Y, Z, SYM, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, betax, betaxx, betaxy, betaxz, X, Y, Z, ANTI, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, betay, betayx, betayy, betayz, X, Y, Z, SYM, ANTI, SYM, Symmetry, Lev);
-    fderivs(ex, betaz, betazx, betazy, betazz, X, Y, Z, SYM, SYM, ANTI, Symmetry, Lev);
-    fderivs(ex, chi, chix, chiy, chiz, X, Y, Z, SYM, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, dxx, gxxx, gxxy, gxxz, X, Y, Z, SYM, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, gxy, gxyx, gxyy, gxyz, X, Y, Z, ANTI, ANTI, SYM, Symmetry, Lev);
-    fderivs(ex, gxz, gxzx, gxzy, gxzz, X, Y, Z, ANTI, SYM, ANTI, Symmetry, Lev);
-    fderivs(ex, dyy, gyyx, gyyy, gyyz, X, Y, Z, SYM, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, gyz, gyzx, gyzy, gyzz, X, Y, Z, SYM, ANTI, ANTI, Symmetry, Lev);
-    fderivs(ex, dzz, gzzx, gzzy, gzzz, X, Y, Z, SYM, SYM, SYM, Symmetry, Lev);
-
-    /* EM field derivatives */
-    fderivs(ex, Kpsi, Kpsix, Kpsiy, Kpsiz, X, Y, Z, SYM, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, Kphi, Kphix, Kphiy, Kphiz, X, Y, Z, SYM, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, Ex, Exx, Exy, Exz, X, Y, Z, ANTI, SYM, SYM, Symmetry, Lev);
-    fderivs(ex, Ey, Eyx, Eyy, Eyz, X, Y, Z, SYM, ANTI, SYM, Symmetry, Lev);
-    fderivs(ex, Ez, Ezx, Ezy, Ezz, X, Y, Z, SYM, SYM, ANTI, Symmetry, Lev);
-    fderivs(ex, Bx, Bxx, Bxy, Bxz, X, Y, Z, SYM, ANTI, ANTI, Symmetry, Lev);
-    fderivs(ex, By, Byx, Byy, Byz, X, Y, Z, ANTI, SYM, ANTI, Symmetry, Lev);
-    fderivs(ex, Bz, Bzx, Bzy, Bzz, X, Y, Z, ANTI, ANTI, SYM, Symmetry, Lev);
-
-    /* ==== 2. Compute EM RHS and stress-energy ==== */
-    const double F1o4PI = ONE / (FOUR * PI);
-    for (size_t i = 0; i < n; ++i) {
-        const double alpn1 = Lap[i] + ONE;
-        const double chin1 = chi[i] + ONE;
-        const double chi3o2 = sqrt(chin1) * chin1;  // chi^{3/2}
-        const double ichi = ONE / chin1;
-
-        /* physical metric */
-        const double pgxx = (dxx[i] + ONE) * ichi;
-        const double pgyy = (dyy[i] + ONE) * ichi;
-        const double pgzz = (dzz[i] + ONE) * ichi;
-        const double pgxy = gxy[i] * ichi;
-        const double pgxz = gxz[i] * ichi;
-        const double pgyz = gyz[i] * ichi;
-
-        /* inverse physical metric */
-        const double det = pgxx * pgyy * pgzz + pgxy * pgyz * pgxz + pgxz * pgxy * pgyz
-                         - pgxz * pgyy * pgxz - pgxy * pgxy * pgzz - pgxx * pgyz * pgyz;
-        const double idet = ONE / det;
-        const double upxx = (pgyy * pgzz - pgyz * pgyz) * idet;
-        const double upxy = -(pgxy * pgzz - pgyz * pgxz) * idet;
-        const double upxz = (pgxy * pgyz - pgyy * pgxz) * idet;
-        const double upyy = (pgxx * pgzz - pgxz * pgxz) * idet;
-        const double upyz = -(pgxx * pgyz - pgxy * pgxz) * idet;
-        const double upzz = (pgxx * pgyy - pgxy * pgxy) * idet;
-        gupxx[i]=upxx; gupxy[i]=upxy; gupxz[i]=upxz;
-        gupyy[i]=upyy; gupyz[i]=upyz; gupzz[i]=upzz;
-
-        /* E-field RHS */
-        /* curl(B) part: epsilon^{ijk} ∂_j (alpha * B_k) in coordinate basis */
-        /* Using lower-index B fields: B_i_lower = pg_{ij} * B^j */
-        const double BxL = pgxx*Bx[i] + pgxy*By[i] + pgxz*Bz[i];
-        const double ByL = pgxy*Bx[i] + pgyy*By[i] + pgyz*Bz[i];
-        const double BzL = pgxz*Bx[i] + pgyz*By[i] + pgzz*Bz[i];
-
-        /* Physical metric derivatives (chain rule from conformal) */
-        const double pgxx_x = (gxxx[i] - pgxx*chix[i]) * ichi;
-        /* const double pgxx_y = (gxxy[i] - pgxx*chiy[i]) * ichi; */
-        const double pgxy_x = (gxyx[i] - pgxy*chix[i]) * ichi;
-        const double pgxy_y = (gxyy[i] - pgxy*chiy[i]) * ichi;
-        const double pgxz_x = (gxzx[i] - pgxz*chix[i]) * ichi;
-        const double pgxz_z = (gxzz[i] - pgxz*chiz[i]) * ichi;
-        const double pgyy_y = (gyyy[i] - pgyy*chiy[i]) * ichi;
-        const double pgyz_y = (gyzy[i] - pgyz*chiy[i]) * ichi;
-        const double pgyz_z = (gyzz[i] - pgyz*chiz[i]) * ichi;
-        const double pgzz_z = (gzzz[i] - pgzz*chiz[i]) * ichi;
-
-        /* Curl_x(B) = ∂_y (alpha*BzL) - ∂_z (alpha*ByL) */
-        const double aBx = alpn1*BxL, aBy = alpn1*ByL, aBz = alpn1*BzL;
-        const double curlBx = (aBz*Lapy[i] + alpn1*(pgxz*Bxy[i]+pgyz*Byy[i]+pgzz*Bzy[i]) + alpn1*(Bx[i]*gxzy[i]+By[i]*gyzy[i]+Bz[i]*gzzy[i]))
-                            - (aBy*Lapz[i] + alpn1*(pgxy*Bxz[i]+pgyy*Byz[i]+pgyz*Bzz[i]) + alpn1*(Bx[i]*gxyz[i]+By[i]*gyyz[i]+Bz[i]*gyzz[i]));
-        double curlBy = (aBx*Lapz[i] + alpn1*(pgxx*Bxz[i]+pgxy*Byz[i]+pgxz*Bzz[i]) + alpn1*(Bx[i]*gxxz[i]+By[i]*gxyz[i]+Bz[i]*gxzz[i]))
-                      - (aBz*Lapx[i] + alpn1*(pgxz*Bxx[i]+pgyz*Byx[i]+pgzz*Bzx[i]) + alpn1*(Bx[i]*gxzx[i]+By[i]*gyzx[i]+Bz[i]*gzzx[i]));
-        double curlBz = (aBy*Lapx[i] + alpn1*(pgxy*Bxx[i]+pgyy*Byx[i]+pgyz*Bzx[i]) + alpn1*(Bx[i]*gxyx[i]+By[i]*gyyx[i]+Bz[i]*gyzx[i]))
-                      - (aBx*Lapy[i] + alpn1*(pgxx*Bxy[i]+pgxy*Byy[i]+pgxz*Bzy[i]) + alpn1*(Bx[i]*gxxy[i]+By[i]*gxyy[i]+Bz[i]*gxzy[i]));
-
-        /* Advection part: -beta^j * ∂_j E^i */
-        const double advEx = Ex[i]*betaxx[i] + Ey[i]*betaxy[i] + Ez[i]*betaxz[i];
-        const double advEy = Ex[i]*betayx[i] + Ey[i]*betayy[i] + Ez[i]*betayz[i];
-        const double advEz = Ex[i]*betazx[i] + Ey[i]*betazy[i] + Ez[i]*betazz[i];
-
-        /* grad(Kpsi) contracted with inverse metric */
-        const double gupKx = upxx*Kpsix[i] + upxy*Kpsiy[i] + upxz*Kpsiz[i];
-        const double gupKy = upxy*Kpsix[i] + upyy*Kpsiy[i] + upyz*Kpsiz[i];
-        const double gupKz = upxz*Kpsix[i] + upyz*Kpsiy[i] + upzz*Kpsiz[i];
-
-        Ex_rhs[i] = alpn1*trK[i]*Ex[i] - advEx - FOUR*PI*alpn1*Jx[i] - alpn1*gupKx + chi3o2*curlBx;
-        Ey_rhs[i] = alpn1*trK[i]*Ey[i] - advEy - FOUR*PI*alpn1*Jy[i] - alpn1*gupKy + chi3o2*curlBy;
-        Ez_rhs[i] = alpn1*trK[i]*Ez[i] - advEz - FOUR*PI*alpn1*Jz[i] - alpn1*gupKz + chi3o2*curlBz;
-
-        /* B-field RHS: similar but with -chi^{3/2} * curl(E) and grad(Kphi) */
-        const double ExL = pgxx*Ex[i] + pgxy*Ey[i] + pgxz*Ez[i];
-        const double EyL = pgxy*Ex[i] + pgyy*Ey[i] + pgyz*Ez[i];
-        const double EzL = pgxz*Ex[i] + pgyz*Ey[i] + pgzz*Ez[i];
-
-        const double aEx = alpn1*ExL, aEy = alpn1*EyL, aEz = alpn1*EzL;
-        const double curlEx = (aEz*Lapy[i] + alpn1*(pgxz*Exy[i]+pgyz*Eyy[i]+pgzz*Ezy[i]) + alpn1*(Ex[i]*gxzy[i]+Ey[i]*gyzy[i]+Ez[i]*gzzy[i]))
-                            - (aEy*Lapz[i] + alpn1*(pgxy*Exz[i]+pgyy*Eyz[i]+pgyz*Ezz[i]) + alpn1*(Ex[i]*gxyz[i]+Ey[i]*gyyz[i]+Ez[i]*gyzz[i]));
-        double curlEy = (aEx*Lapz[i] + alpn1*(pgxx*Exz[i]+pgxy*Eyz[i]+pgxz*Ezz[i]) + alpn1*(Ex[i]*gxxz[i]+Ey[i]*gxyz[i]+Ez[i]*gxzz[i]))
-                      - (aEz*Lapx[i] + alpn1*(pgxz*Exx[i]+pgyz*Eyx[i]+pgzz*Ezx[i]) + alpn1*(Ex[i]*gxzx[i]+Ey[i]*gyzx[i]+Ez[i]*gzzx[i]));
-        double curlEz = (aEy*Lapx[i] + alpn1*(pgxy*Exx[i]+pgyy*Eyx[i]+pgyz*Ezx[i]) + alpn1*(Ex[i]*gxyx[i]+Ey[i]*gyyx[i]+Ez[i]*gyzx[i]))
-                      - (aEx*Lapy[i] + alpn1*(pgxx*Exy[i]+pgxy*Eyy[i]+pgxz*Ezy[i]) + alpn1*(Ex[i]*gxxy[i]+Ey[i]*gxyy[i]+Ez[i]*gxzy[i]));
-
-        const double advBx = Bx[i]*betaxx[i] + By[i]*betaxy[i] + Bz[i]*betaxz[i];
-        const double advBy = Bx[i]*betayx[i] + By[i]*betayy[i] + Bz[i]*betayz[i];
-        const double advBz = Bx[i]*betazx[i] + By[i]*betazy[i] + Bz[i]*betazz[i];
-
-        const double gupKphix = upxx*Kphix[i] + upxy*Kphiy[i] + upxz*Kphiz[i];
-        const double gupKphiy = upxy*Kphix[i] + upyy*Kphiy[i] + upyz*Kphiz[i];
-        const double gupKphiz = upxz*Kphix[i] + upyz*Kphiy[i] + upzz*Kphiz[i];
-
-        Bx_rhs[i] = alpn1*trK[i]*Bx[i] - advBx - alpn1*gupKphix - chi3o2*curlEx;
-        By_rhs[i] = alpn1*trK[i]*By[i] - advBy - alpn1*gupKphiy - chi3o2*curlEy;
-        Bz_rhs[i] = alpn1*trK[i]*Bz[i] - advBz - alpn1*gupKphiz - chi3o2*curlEz;
-
-        /* Scalar potential RHS */
-        const double divE = Exx[i] + Eyy[i] + Ezz[i];
-        const double divB = Bxx[i] + Byy[i] + Bzz[i];
-        const double chiCont = F3o2 * ichi * (chix[i]*Ex[i] + chiy[i]*Ey[i] + chiz[i]*Ez[i]);
-        Kpsi_rhs[i] = FOUR*PI*alpn1*qchar[i] - alpn1*kappa*Kpsi[i] - alpn1*(divE - chiCont);
-        Kphi_rhs[i] = -alpn1*kappa*Kphi[i] - alpn1*(divB - F3o2*ichi*(chix[i]*Bx[i] + chiy[i]*By[i] + chiz[i]*Bz[i]));
-
-        /* Stress-energy tensor */
-        const double E2 = pgxx*Ex[i]*Ex[i] + pgyy*Ey[i]*Ey[i] + pgzz*Ez[i]*Ez[i]
-                        + TWO*(pgxy*Ex[i]*Ey[i] + pgxz*Ex[i]*Ez[i] + pgyz*Ey[i]*Ez[i]);
-        const double B2 = pgxx*Bx[i]*Bx[i] + pgyy*By[i]*By[i] + pgzz*Bz[i]*Bz[i]
-                        + TWO*(pgxy*Bx[i]*By[i] + pgxz*Bx[i]*Bz[i] + pgyz*By[i]*Bz[i]);
-        rho[i] = (E2 + B2) / (EIT * PI);
-        const double ichi3o2 = ONE / chi3o2;
-        Sx[i] = (Ey[i]*Bz[i] - Ez[i]*By[i]) * F1o4PI * ichi3o2;
-        Sy[i] = (Ez[i]*Bx[i] - Ex[i]*Bz[i]) * F1o4PI * ichi3o2;
-        Sz[i] = (Ex[i]*By[i] - Ey[i]*Bx[i]) * F1o4PI * ichi3o2;
-        const double lExi = pgxx*Ex[i] + pgxy*Ey[i] + pgxz*Ez[i];
-        const double lEyi = pgxy*Ex[i] + pgyy*Ey[i] + pgyz*Ez[i];
-        const double lEzi = pgxz*Ex[i] + pgyz*Ey[i] + pgzz*Ez[i];
-        const double lBxi = pgxx*Bx[i] + pgxy*By[i] + pgxz*Bz[i];
-        const double lByi = pgxy*Bx[i] + pgyy*By[i] + pgyz*Bz[i];
-        const double lBzi = pgxz*Bx[i] + pgyz*By[i] + pgzz*Bz[i];
-        Sxx[i] = rho[i]*pgxx - (lExi*lExi + lBxi*lBxi) * F1o4PI;
-        Sxy[i] = rho[i]*pgxy - (lExi*lEyi + lBxi*lByi) * F1o4PI;
-        Sxz[i] = rho[i]*pgxz - (lExi*lEzi + lBxi*lBzi) * F1o4PI;
-        Syy[i] = rho[i]*pgyy - (lEyi*lEyi + lByi*lByi) * F1o4PI;
-        Syz[i] = rho[i]*pgyz - (lEyi*lEzi + lByi*lBzi) * F1o4PI;
-        Szz[i] = rho[i]*pgzz - (lEzi*lEzi + lBzi*lBzi) * F1o4PI;
-    }
-
-    /* ==== 3. Call BSSN RHS with EM stress-energy ==== */
-    gont = f_compute_rhs_bssn(ex, T, X, Y, Z,
-        chi, trK, dxx, gxy, gxz, dyy, gyz, dzz,
-        Axx, Axy, Axz, Ayy, Ayz, Azz,
-        Gamx, Gamy, Gamz, Lap, betax, betay, betaz, dtSfx, dtSfy, dtSfz,
-        chi_rhs, trK_rhs,
-        gxx_rhs, gxy_rhs, gxz_rhs, gyy_rhs, gyz_rhs, gzz_rhs,
-        Axx_rhs, Axy_rhs, Axz_rhs, Ayy_rhs, Ayz_rhs, Azz_rhs,
-        Gamx_rhs, Gamy_rhs, Gamz_rhs, Lap_rhs, betax_rhs, betay_rhs, betaz_rhs,
-        dtSfx_rhs, dtSfy_rhs, dtSfz_rhs,
-        rho, Sx, Sy, Sz, Sxx, Sxy, Sxz, Syy, Syz, Szz,
-        Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-        Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-        Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-        Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-        ham_Res, movx_Res, movy_Res, movz_Res,
-        Gmx_Res, Gmy_Res, Gmz_Res,
-        Symmetry, Lev, eps, co);
-    if (!gont) {
-
-    /* ==== 4. Advection terms for EM fields ==== */
-    lopsided(ex, X, Y, Z, Kpsi, Kpsi_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, Kphi, Kphi_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, Ex,   Ex_rhs,   betax, betay, betaz, Symmetry, ASS);
-    lopsided(ex, X, Y, Z, Ey,   Ey_rhs,   betax, betay, betaz, Symmetry, SAS);
-    lopsided(ex, X, Y, Z, Ez,   Ez_rhs,   betax, betay, betaz, Symmetry, SSA);
-    lopsided(ex, X, Y, Z, Bx,   Bx_rhs,   betax, betay, betaz, Symmetry, SAA);
-    lopsided(ex, X, Y, Z, By,   By_rhs,   betax, betay, betaz, Symmetry, ASA);
-    lopsided(ex, X, Y, Z, Bz,   Bz_rhs,   betax, betay, betaz, Symmetry, AAS);
-
-    /* ==== 5. KO dissipation for EM fields ==== */
-    if (eps > ZEO) {
-        kodis(ex, X, Y, Z, Kpsi, Kpsi_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Kphi, Kphi_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Ex,   Ex_rhs,   ASS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Ey,   Ey_rhs,   SAS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Ez,   Ez_rhs,   SSA, Symmetry, eps);
-        kodis(ex, X, Y, Z, Bx,   Bx_rhs,   SAA, Symmetry, eps);
-        kodis(ex, X, Y, Z, By,   By_rhs,   ASA, Symmetry, eps);
-        kodis(ex, X, Y, Z, Bz,   Bz_rhs,   AAS, Symmetry, eps);
-    }
-
-    /* ==== 6. NaN check ==== */
-        for (int i = 0; i < all; ++i) {
-            if (!isfinite(Ex_rhs[i]+Ey_rhs[i]+Ez_rhs[i]+Bx_rhs[i]+By_rhs[i]+Bz_rhs[i]+Kpsi_rhs[i]+Kphi_rhs[i])) {
-                gont = 1; break;
-            }
-        }
-        } /* inner if (!gont) */
-    } /* outer if (!gont) */
-
-    free(chix);free(chiy);free(chiz);
-    free(Exx);free(Exy);free(Exz);free(Eyx);free(Eyy);free(Eyz);free(Ezx);free(Ezy);free(Ezz);
-    free(Bxx);free(Bxy);free(Bxz);free(Byx);free(Byy);free(Byz);free(Bzx);free(Bzy);free(Bzz);
-    free(Kpsix);free(Kpsiy);free(Kpsiz);
-    free(Kphix);free(Kphiy);free(Kphiz);
-    free(Lapx);free(Lapy);free(Lapz);
-    free(betaxx);free(betaxy);free(betaxz);free(betayx);free(betayy);free(betayz);free(betazx);free(betazy);free(betazz);
-    free(gxxx);free(gxxy);free(gxxz);free(gxyx);free(gxyy);free(gxyz);free(gxzx);free(gxzy);free(gxzz);
-    free(gyyx);free(gyyy);free(gyyz);free(gyzx);free(gyzy);free(gyzz);free(gzzx);free(gzzy);free(gzzz);
-    free(gupxx);free(gupxy);free(gupxz);free(gupyy);free(gupyz);free(gupzz);
-    return gont;
-}

AMSS_NCKU_source/bssn_escalar_rhs_c.C

@@ -1,169 +0,0 @@
-#include "macrodef.h"
-#include "bssn_rhs.h"
-#include "share_func.h"
-#include "tool.h"
-#include <vector>
-
-namespace
-{
-    // Reuse the temporary workspace across block calls to avoid repeated heap churn
-    // in the EScalar wrapper. MPI ranks execute this path sequentially, so a single
-    // process-local buffer is sufficient here.
-    std::vector<double> g_escalar_tmp_store;
-}
-
-#ifdef fortran1
-#define f_frpotential frpotential
-#endif
-#ifdef fortran2
-#define f_frpotential FRPOTENTIAL
-#endif
-#ifdef fortran3
-#define f_frpotential frpotential_
-#endif
-
-extern "C"
-{
-    void f_frpotential(int *, double *, double *, double *);
-}
-
-int f_compute_rhs_bssn_escalar_c(int *ex, double &T,
-                                 double *X, double *Y, double *Z,
-                                 double *chi, double *trK,
-                                 double *dxx, double *gxy, double *gxz, double *dyy, double *gyz, double *dzz,
-                                 double *Axx, double *Axy, double *Axz, double *Ayy, double *Ayz, double *Azz,
-                                 double *Gamx, double *Gamy, double *Gamz,
-                                 double *Lap, double *betax, double *betay, double *betaz,
-                                 double *dtSfx, double *dtSfy, double *dtSfz,
-                                 double *Sphi, double *Spi,
-                                 double *chi_rhs, double *trK_rhs,
-                                 double *gxx_rhs, double *gxy_rhs, double *gxz_rhs, double *gyy_rhs, double *gyz_rhs, double *gzz_rhs,
-                                 double *Axx_rhs, double *Axy_rhs, double *Axz_rhs, double *Ayy_rhs, double *Ayz_rhs, double *Azz_rhs,
-                                 double *Gamx_rhs, double *Gamy_rhs, double *Gamz_rhs,
-                                 double *Lap_rhs, double *betax_rhs, double *betay_rhs, double *betaz_rhs,
-                                 double *dtSfx_rhs, double *dtSfy_rhs, double *dtSfz_rhs,
-                                 double *Sphi_rhs, double *Spi_rhs,
-                                 double *rho, double *Sx, double *Sy, double *Sz,
-                                 double *Sxx, double *Sxy, double *Sxz, double *Syy, double *Syz, double *Szz,
-                                 double *Gamxxx, double *Gamxxy, double *Gamxxz, double *Gamxyy, double *Gamxyz, double *Gamxzz,
-                                 double *Gamyxx, double *Gamyxy, double *Gamyxz, double *Gamyyy, double *Gamyyz, double *Gamyzz,
-                                 double *Gamzxx, double *Gamzxy, double *Gamzxz, double *Gamzyy, double *Gamzyz, double *Gamzzz,
-                                 double *Rxx, double *Rxy, double *Rxz, double *Ryy, double *Ryz, double *Rzz,
-                                 double *ham_Res, double *movx_Res, double *movy_Res, double *movz_Res,
-                                 double *Gmx_Res, double *Gmy_Res, double *Gmz_Res,
-                                 int &Symmetry, int &Lev, double &eps, int &co)
-{
-    const int nx = ex[0], ny = ex[1], nz = ex[2];
-    const int all = nx * ny * nz;
-
-    const size_t workspace_size = size_t(all) * 17;
-    if (g_escalar_tmp_store.size() < workspace_size)
-        g_escalar_tmp_store.resize(workspace_size);
-
-    double *tmp_ptr = g_escalar_tmp_store.data();
-    auto alloc_tmp = [&](int n = 1) -> double *
-    {
-        double *ptr = tmp_ptr;
-        tmp_ptr += size_t(all) * n;
-        return ptr;
-    };
-
-    double *chix = alloc_tmp(), *chiy = alloc_tmp(), *chiz = alloc_tmp();
-    double *Kx = alloc_tmp(), *Ky = alloc_tmp(), *Kz = alloc_tmp();
-    double *fxx = alloc_tmp(), *fxy = alloc_tmp(), *fxz = alloc_tmp();
-    double *fyy = alloc_tmp(), *fyz = alloc_tmp(), *fzz = alloc_tmp();
-    double *Lapx = alloc_tmp(), *Lapy = alloc_tmp(), *Lapz = alloc_tmp();
-    double *V = alloc_tmp(), *dVdSphi = alloc_tmp();
-
-    const double ZEO = 0.0, ONE = 1.0, TWO = 2.0, HALF = 0.5;
-    const double SSS[3] = {1.0, 1.0, 1.0};
-
-    fderivs(ex, chi, chix, chiy, chiz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, Lap, Lapx, Lapy, Lapz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, Sphi, Kx, Ky, Kz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, Sphi, fxx, fxy, fxz, fyy, fyz, fzz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-
-    f_frpotential(ex, Sphi, V, dVdSphi);
-
-    for (int i = 0; i < all; ++i)
-    {
-        const double alpn1 = Lap[i] + ONE;
-        const double chin1 = chi[i] + ONE;
-        const double gxx = dxx[i] + ONE;
-        const double gyy = dyy[i] + ONE;
-        const double gzz = dzz[i] + ONE;
-        const double det = gxx * gyy * gzz + gxy[i] * gyz[i] * gxz[i] + gxz[i] * gxy[i] * gyz[i]
-                         - gxz[i] * gyy * gxz[i] - gxy[i] * gxy[i] * gzz - gxx * gyz[i] * gyz[i];
-        const double gupxx = (gyy * gzz - gyz[i] * gyz[i]) / det;
-        const double gupxy = -(gxy[i] * gzz - gyz[i] * gxz[i]) / det;
-        const double gupxz = (gxy[i] * gyz[i] - gyy * gxz[i]) / det;
-        const double gupyy = (gxx * gzz - gxz[i] * gxz[i]) / det;
-        const double gupyz = -(gxx * gyz[i] - gxy[i] * gxz[i]) / det;
-        const double gupzz = (gxx * gyy - gxy[i] * gxy[i]) / det;
-
-        Sphi_rhs[i] = alpn1 * Spi[i];
-
-        Spi_rhs[i] = gupxx * fxx[i] + gupyy * fyy[i] + gupzz * fzz[i]
-                   + TWO * (gupxy * fxy[i] + gupxz * fxz[i] + gupyz * fyz[i])
-                   - ((Gamx[i] + (gupxx * chix[i] + gupxy * chiy[i] + gupxz * chiz[i]) / TWO / chin1) * Kx[i]
-                   +  (Gamy[i] + (gupxy * chix[i] + gupyy * chiy[i] + gupyz * chiz[i]) / TWO / chin1) * Ky[i]
-                   +  (Gamz[i] + (gupxz * chix[i] + gupyz * chiy[i] + gupzz * chiz[i]) / TWO / chin1) * Kz[i]);
-
-        Spi_rhs[i] = Spi_rhs[i] * alpn1
-                   + gupxx * Lapx[i] * Kx[i] + gupxy * Lapx[i] * Ky[i] + gupxz * Lapx[i] * Kz[i]
-                   + gupxy * Lapy[i] * Kx[i] + gupyy * Lapy[i] * Ky[i] + gupyz * Lapy[i] * Kz[i]
-                   + gupxz * Lapz[i] * Kx[i] + gupyz * Lapz[i] * Ky[i] + gupzz * Lapz[i] * Kz[i];
-
-        Spi_rhs[i] = Spi_rhs[i] * chin1 + alpn1 * (trK[i] * Spi[i] - dVdSphi[i]);
-
-        rho[i] = chin1 * ((gupxx * Kx[i] * Kx[i] + gupyy * Ky[i] * Ky[i] + gupzz * Kz[i] * Kz[i]) * HALF
-               + gupxy * Kx[i] * Ky[i] + gupxz * Kx[i] * Kz[i] + gupyz * Ky[i] * Kz[i])
-               + Spi[i] * Spi[i] * HALF + V[i];
-        Sx[i] = -Spi[i] * Kx[i];
-        Sy[i] = -Spi[i] * Ky[i];
-        Sz[i] = -Spi[i] * Kz[i];
-
-        const double pressure = (rho[i] - Spi[i] * Spi[i]) / chin1;
-        Sxx[i] = Kx[i] * Kx[i] - pressure * gxx;
-        Sxy[i] = Kx[i] * Ky[i] - pressure * gxy[i];
-        Sxz[i] = Kx[i] * Kz[i] - pressure * gxz[i];
-        Syy[i] = Ky[i] * Ky[i] - pressure * gyy;
-        Syz[i] = Ky[i] * Kz[i] - pressure * gyz[i];
-        Szz[i] = Kz[i] * Kz[i] - pressure * gzz;
-    }
-
-    if (f_compute_rhs_bssn(ex, T, X, Y, Z,
-                           chi, trK,
-                           dxx, gxy, gxz, dyy, gyz, dzz,
-                           Axx, Axy, Axz, Ayy, Ayz, Azz,
-                           Gamx, Gamy, Gamz,
-                           Lap, betax, betay, betaz,
-                           dtSfx, dtSfy, dtSfz,
-                           chi_rhs, trK_rhs,
-                           gxx_rhs, gxy_rhs, gxz_rhs, gyy_rhs, gyz_rhs, gzz_rhs,
-                           Axx_rhs, Axy_rhs, Axz_rhs, Ayy_rhs, Ayz_rhs, Azz_rhs,
-                           Gamx_rhs, Gamy_rhs, Gamz_rhs,
-                           Lap_rhs, betax_rhs, betay_rhs, betaz_rhs,
-                           dtSfx_rhs, dtSfy_rhs, dtSfz_rhs,
-                           rho, Sx, Sy, Sz,
-                           Sxx, Sxy, Sxz, Syy, Syz, Szz,
-                           Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-                           Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-                           Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-                           Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-                           ham_Res, movx_Res, movy_Res, movz_Res,
-                           Gmx_Res, Gmy_Res, Gmz_Res,
-                           Symmetry, Lev, eps, co))
-        return 1;
-
-    lopsided_kodis(ex, X, Y, Z, Sphi, Sphi_rhs, betax, betay, betaz, Symmetry, SSS, eps);
-    lopsided_kodis(ex, X, Y, Z, Spi, Spi_rhs, betax, betay, betaz, Symmetry, SSS, eps);
-
-    for (int i = 0; i < all; ++i)
-    {
-        if (Sphi_rhs[i] != Sphi_rhs[i] || Spi_rhs[i] != Spi_rhs[i] || rho[i] != rho[i])
-            return 1;
-    }
-
-    return 0;
-}

AMSS_NCKU_source/bssn_gpu.h

@@ -4,10 +4,8 @@
 #include "bssn_macro.h"
 #include "macrodef.fh"
 
-#define DEVICE_ID 0
-// #define DEVICE_ID_BY_MPI_RANK
-#define GRID_DIM 256
-#define BLOCK_DIM 128
+#define GRID_DIM 256
+#define BLOCK_DIM 128
 
 #define _FH2_(i, j, k) fh[(i) + (j) * _1D_SIZE[2] + (k) * _2D_SIZE[2]]
 #define _FH3_(i, j, k) fh[(i) + (j) * _1D_SIZE[3] + (k) * _2D_SIZE[3]]
@@ -65,9 +63,45 @@ int gpu_rhs(int calledby, int mpi_rank, int *ex, double &T,
             double *Gmx_Res, double *Gmy_Res, double *Gmz_Res,
             int &Symmetry, int &Lev, double &eps, int &co);
 
-int gpu_rhs_ss(RHS_SS_PARA);
-
-/** Init GPU side data in GPUMeta. */
-// void init_fluid_meta_gpu(GPUMeta *gpu_meta);
+int gpu_rhs_ss(RHS_SS_PARA);
+
+int bssn_gpu_bind_process_device(int mpi_rank);
+void bssn_gpu_clear_cached_device_buffers();
+void bssn_gpu_release_pinned_host_buffers();
+const double *bssn_gpu_find_device_buffer(const double *host_ptr);
+void bssn_gpu_register_device_buffer(const double *host_ptr, const double *device_ptr);
+void bssn_gpu_prepare_host_buffer(const double *host_ptr, int count);
+int bssn_gpu_stage_upload_buffer(const double *host_ptr, int count);
+int bssn_gpu_stage_zero_buffer(const double *host_ptr, int count);
+int bssn_gpu_stage_upload_region(const double *host_ptr,
+                                 const int *full_shape,
+                                 const double *full_llb,
+                                 const double *full_uub,
+                                 const int *region_shape,
+                                 const double *region_llb);
+int bssn_gpu_stage_download_region(double *host_ptr,
+                                   const int *full_shape,
+                                   const double *full_llb,
+                                   const double *full_uub,
+                                   const int *region_shape,
+                                   const double *region_llb);
+int bssn_gpu_stage_download_region_to_buffer(const double *host_src_ptr,
+                                             const int *full_shape,
+                                             const double *full_llb,
+                                             const double *full_uub,
+                                             const int *region_shape,
+                                             const double *region_llb,
+                                             double *host_dst_ptr);
+int bssn_gpu_stage_upload_buffer_to_region(const double *host_src_ptr,
+                                           double *host_dst_ptr,
+                                           const int *full_shape,
+                                           const double *full_llb,
+                                           const double *full_uub,
+                                           const int *region_shape,
+                                           const double *region_llb);
+int bssn_gpu_download_buffer_batch(const int *ex, double **host_ptrs, int num_buffers);
+
+/** Init GPU side data in GPUMeta. */
+// void init_fluid_meta_gpu(GPUMeta *gpu_meta);
 
 #endif

AMSS_NCKU_source/bssn_macro.h

@@ -65,9 +65,10 @@ if(TIME_COUNT_EACH_RANK == 1){\
      }\
 }
 
-//3---------------------GPU---------------------
-#define CALLED_BY_STEP 0
-#define CALLED_BY_CONSTRAINT 1
+//3---------------------GPU---------------------
+#define CALLED_BY_STEP 0
+#define CALLED_BY_CONSTRAINT 1
+#define CALLED_BY_CONSTRAINT_CONS_ONLY 2
 
 #define RHS_PARA_CALLED_FIRST_TIME cg->shape,TRK4,cg->X[0],cg->X[1],cg->X[2],cg->fgfs[phi0->sgfn],cg->fgfs[trK0->sgfn],cg->fgfs[gxx0->sgfn],cg->fgfs[gxy0->sgfn],cg->fgfs[gxz0->sgfn],cg->fgfs[gyy0->sgfn],cg->fgfs[gyz0->sgfn],cg->fgfs[gzz0->sgfn],cg->fgfs[Axx0->sgfn],cg->fgfs[Axy0->sgfn],cg->fgfs[Axz0->sgfn],cg->fgfs[Ayy0->sgfn],cg->fgfs[Ayz0->sgfn],cg->fgfs[Azz0->sgfn],cg->fgfs[Gmx0->sgfn],cg->fgfs[Gmy0->sgfn],cg->fgfs[Gmz0->sgfn],cg->fgfs[Lap0->sgfn],cg->fgfs[Sfx0->sgfn],cg->fgfs[Sfy0->sgfn],cg->fgfs[Sfz0->sgfn],cg->fgfs[dtSfx0->sgfn],cg->fgfs[dtSfy0->sgfn],cg->fgfs[dtSfz0->sgfn],cg->fgfs[phi_rhs->sgfn],cg->fgfs[trK_rhs->sgfn],cg->fgfs[gxx_rhs->sgfn],cg->fgfs[gxy_rhs->sgfn],cg->fgfs[gxz_rhs->sgfn],cg->fgfs[gyy_rhs->sgfn],cg->fgfs[gyz_rhs->sgfn],cg->fgfs[gzz_rhs->sgfn],cg->fgfs[Axx_rhs->sgfn],cg->fgfs[Axy_rhs->sgfn],cg->fgfs[Axz_rhs->sgfn],cg->fgfs[Ayy_rhs->sgfn],cg->fgfs[Ayz_rhs->sgfn],cg->fgfs[Azz_rhs->sgfn],cg->fgfs[Gmx_rhs->sgfn],cg->fgfs[Gmy_rhs->sgfn],cg->fgfs[Gmz_rhs->sgfn],cg->fgfs[Lap_rhs->sgfn],cg->fgfs[Sfx_rhs->sgfn],cg->fgfs[Sfy_rhs->sgfn],cg->fgfs[Sfz_rhs->sgfn],cg->fgfs[dtSfx_rhs->sgfn],cg->fgfs[dtSfy_rhs->sgfn],cg->fgfs[dtSfz_rhs->sgfn],cg->fgfs[rho->sgfn],cg->fgfs[Sx->sgfn],cg->fgfs[Sy->sgfn],cg->fgfs[Sz->sgfn],cg->fgfs[Sxx->sgfn],cg->fgfs[Sxy->sgfn],cg->fgfs[Sxz->sgfn],cg->fgfs[Syy->sgfn],cg->fgfs[Syz->sgfn],cg->fgfs[Szz->sgfn],cg->fgfs[Gamxxx->sgfn],cg->fgfs[Gamxxy->sgfn],cg->fgfs[Gamxxz->sgfn],cg->fgfs[Gamxyy->sgfn],cg->fgfs[Gamxyz->sgfn],cg->fgfs[Gamxzz->sgfn],cg->fgfs[Gamyxx->sgfn],cg->fgfs[Gamyxy->sgfn],cg->fgfs[Gamyxz->sgfn],cg->fgfs[Gamyyy->sgfn],cg->fgfs[Gamyyz->sgfn],cg->fgfs[Gamyzz->sgfn],cg->fgfs[Gamzxx->sgfn],cg->fgfs[Gamzxy->sgfn],cg->fgfs[Gamzxz->sgfn],cg->fgfs[Gamzyy->sgfn],cg->fgfs[Gamzyz->sgfn],cg->fgfs[Gamzzz->sgfn],cg->fgfs[Rxx->sgfn],cg->fgfs[Rxy->sgfn],cg->fgfs[Rxz->sgfn],cg->fgfs[Ryy->sgfn],cg->fgfs[Ryz->sgfn],cg->fgfs[Rzz->sgfn],cg->fgfs[Cons_Ham->sgfn],cg->fgfs[Cons_Px->sgfn],cg->fgfs[Cons_Py->sgfn],cg->fgfs[Cons_Pz->sgfn],cg->fgfs[Cons_Gx->sgfn],cg->fgfs[Cons_Gy->sgfn],cg->fgfs[Cons_Gz->sgfn],Symmetry,lev,ndeps,pre
 

AMSS_NCKU_source/bssn_rhs.h

@@ -32,19 +32,6 @@
 #define f_compute_rhs_Z4c_ss compute_rhs_z4c_ss_
 #define f_compute_constraint_fr compute_constraint_fr_
 #endif
-
-#ifdef __cplusplus
-extern "C"
-{
-#endif
-        void f_bssn_rhs_kernel_timing_reset();
-        int f_bssn_rhs_kernel_timing_bucket_count();
-        const double *f_bssn_rhs_kernel_timing_local_seconds();
-        const char *f_bssn_rhs_kernel_timing_label(int);
-#ifdef __cplusplus
-}
-#endif
-
 extern "C"
 {
         int f_compute_rhs_bssn(int *, double &, double *, double *, double *,                                                      // ex,T,X,Y,Z
@@ -67,27 +54,6 @@ extern "C"
                                int &, int &, double &, int &);
 }
 
-int f_compute_rhs_bssn_escalar_c(int *, double &, double *, double *, double *,                                                      // ex,T,X,Y,Z
-                                 double *, double *,                                                                                 // chi, trK
-                                 double *, double *, double *, double *, double *, double *,                                         // gij
-                                 double *, double *, double *, double *, double *, double *,                                         // Aij
-                                 double *, double *, double *,                                                                       // Gam
-                                 double *, double *, double *, double *, double *, double *, double *,                               // Gauge
-                                 double *, double *,                                                                                 // Sphi, Spi
-                                 double *, double *,                                                                                 // chi, trK
-                                 double *, double *, double *, double *, double *, double *,                                         // gij
-                                 double *, double *, double *, double *, double *, double *,                                         // Aij
-                                 double *, double *, double *,                                                                       // Gam
-                                 double *, double *, double *, double *, double *, double *, double *,                               // Gauge
-                                 double *, double *,                                                                                 // Sphi, Spi
-                                 double *, double *, double *, double *, double *, double *, double *, double *, double *, double *, // stress-energy
-                                 double *, double *, double *, double *, double *, double *,                                         // Christoffel
-                                 double *, double *, double *, double *, double *, double *,                                         // Christoffel
-                                 double *, double *, double *, double *, double *, double *,                                         // Christoffel
-                                 double *, double *, double *, double *, double *, double *,                                         // Ricci
-                                 double *, double *, double *, double *, double *, double *, double *,                               // constraint violation
-                                 int &, int &, double &, int &);
-
 extern "C"
 {
         int f_compute_rhs_bssn_ss(int *, double &, double *, double *, double *,                                                      // ex,T,rho,sigma,R
@@ -262,31 +228,4 @@ extern "C"
                                      double *);
 } // FR_cons
 
-// BSSN-EM C kernel (replaces empart.f90 + bssn_rhs.f90 for BSSN+Maxwell)
-int f_compute_rhs_bssn_em_c(int *, double &, double *, double *, double *,
-                             double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *,
-                             double *, double *,
-                             double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *,
-                             double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             double *, double *, double *, double *, double *, double *,
-                             int &, int &, double &, int &);
-
 #endif /* BSSN_H */

AMSS_NCKU_source/bssn_rhs_c.C

@@ -2,88 +2,12 @@
 #include "bssn_rhs.h"
 #include "share_func.h"
 #include "tool.h"
-#include <time.h>
 // 0-based i,j,k
 // #define IDX_F(i,j,k,nx,ny) ((i) + (j)*(nx) + (k)*(nx)*(ny))
 // ex(1)=nx, ex(2)=ny, ex(3)=nz
 
 // 用法：a[ IDX_F(i,j,k,nx,ny) ]
 
-#ifndef BSSN_KERNEL_FINE_TIMING
-#define BSSN_KERNEL_FINE_TIMING 0
-#endif
-
-#if BSSN_KERNEL_FINE_TIMING
-namespace rhs_kernel_timing
-{
-    enum Bucket
-    {
-        KB_SETUP_DERIVS = 0,
-        KB_GEOM_GAMMA,
-        KB_RICCI_METRIC,
-        KB_CHI_LAPSE,
-        KB_AIJ_TRK_GAUGE,
-        KB_KO_CONSTRAINT,
-        KB_COUNT
-    };
-
-    static double local_bucket_seconds[KB_COUNT];
-
-    static const char *bucket_labels[KB_COUNT] =
-    {
-        "setup_derivs",
-        "geom_gamma",
-        "ricci_metric",
-        "chi_lapse",
-        "aij_trk_gauge",
-        "ko_constraint"
-    };
-
-    static inline double now_seconds()
-    {
-        struct timespec ts;
-        clock_gettime(CLOCK_MONOTONIC, &ts);
-        return double(ts.tv_sec) + 1.0e-9 * double(ts.tv_nsec);
-    }
-}
-
-extern "C" void f_bssn_rhs_kernel_timing_reset()
-{
-    for (int i = 0; i < rhs_kernel_timing::KB_COUNT; ++i)
-        rhs_kernel_timing::local_bucket_seconds[i] = 0.0;
-}
-
-extern "C" int f_bssn_rhs_kernel_timing_bucket_count()
-{
-    return rhs_kernel_timing::KB_COUNT;
-}
-
-extern "C" const double *f_bssn_rhs_kernel_timing_local_seconds()
-{
-    return rhs_kernel_timing::local_bucket_seconds;
-}
-
-extern "C" const char *f_bssn_rhs_kernel_timing_label(int bucket_index)
-{
-    if (bucket_index < 0 || bucket_index >= rhs_kernel_timing::KB_COUNT)
-        return "unknown";
-    return rhs_kernel_timing::bucket_labels[bucket_index];
-}
-
-#define RHS_KERNEL_TIMER_DECL(var_name) const double var_name = rhs_kernel_timing::now_seconds()
-#define RHS_KERNEL_TIMER_ADD(bucket_name, var_name) \
-    rhs_kernel_timing::local_bucket_seconds[int(rhs_kernel_timing::bucket_name)] += \
-        rhs_kernel_timing::now_seconds() - (var_name)
-#else
-extern "C" void f_bssn_rhs_kernel_timing_reset() {}
-extern "C" int f_bssn_rhs_kernel_timing_bucket_count() { return 0; }
-extern "C" const double *f_bssn_rhs_kernel_timing_local_seconds() { return 0; }
-extern "C" const char *f_bssn_rhs_kernel_timing_label(int) { return "disabled"; }
-
-#define RHS_KERNEL_TIMER_DECL(var_name)
-#define RHS_KERNEL_TIMER_ADD(bucket_name, var_name)
-#endif
-
 // C function that calculates the right-hand side for BSSN equations
 int f_compute_rhs_bssn(int *ex, double &T, 
                        double *X, double *Y, double *Z,
@@ -178,7 +102,6 @@ int f_compute_rhs_bssn(int *ex, double &T,
     dY = Y[1] - Y[0];
     dZ = Z[1] - Z[0];
 
-        RHS_KERNEL_TIMER_DECL(timer_setup_derivs);
         // 1ms //
         for(int i=0;i<all;i+=1){
             alpn1[i] = Lap[i] + 1.0;
@@ -218,8 +141,6 @@ int f_compute_rhs_bssn(int *ex, double &T,
                         (dxx[i] + ONE) * betaxz[i] + gxy[i] * betayz[i] + gyz[i] * betayx[i] 
                         + (dzz[i] + ONE) * betazx[i] - gxz[i] * betayy[i];                  
         }
-        RHS_KERNEL_TIMER_ADD(KB_SETUP_DERIVS, timer_setup_derivs);
-        RHS_KERNEL_TIMER_DECL(timer_geom_gamma);
         // Fused: inverse metric + Gamma constraint + Christoffel (3 loops -> 1)
         for(int i=0;i<all;i+=1){
             double det = (dxx[i] + ONE) * (dyy[i] + ONE) * (dzz[i] + ONE) + gxy[i] * gyz[i] * gxz[i] + gxz[i] * gxy[i] * gyz[i] -
@@ -362,6 +283,9 @@ int f_compute_rhs_bssn(int *ex, double &T,
                 + ( gupxy[i]*gupyz[i] + gupyy[i]*gupxz[i] ) * Axy[i]
                 + ( gupxy[i]*gupzz[i] + gupyz[i]*gupxz[i] ) * Axz[i]
                 + ( gupyy[i]*gupzz[i] + gupyz[i]*gupyz[i] ) * Ayz[i];
+            Rxx[i] = axx; Ryy[i] = ayy; Rzz[i] = azz;
+            Rxy[i] = axy; Rxz[i] = axz; Ryz[i] = ayz;
+
             Gamx_rhs[i] = - TWO * ( Lapx[i]*axx + Lapy[i]*axy + Lapz[i]*axz ) +
                 TWO * alpn1[i] * (
                 -F3o2/chin1[i] * ( chix[i]*axx + chiy[i]*axy + chiz[i]*axz ) -
@@ -391,8 +315,6 @@ int f_compute_rhs_bssn(int *ex, double &T,
                             + TWO * ( Gamzxy[i]*axy + Gamzxz[i]*axz + Gamzyz[i]*ayz )
                         );
         }
-        RHS_KERNEL_TIMER_ADD(KB_GEOM_GAMMA, timer_geom_gamma);
-        RHS_KERNEL_TIMER_DECL(timer_ricci_metric);
         // 22.3ms //
         fdderivs(ex,betax,gxxx,gxyx,gxzx,gyyx,gyzx,gzzx,
         X,Y,Z,ANTI,SYM, SYM ,Symmetry,Lev);
@@ -410,6 +332,7 @@ int f_compute_rhs_bssn(int *ex, double &T,
             double lfxx = gxxx[i] + gxyy[i] + gxzz[i];
             double lfxy = gxyx[i] + gyyy[i] + gyzz[i];
             double lfxz = gxzx[i] + gyzy[i] + gzzz[i];
+            fxx[i] = lfxx; fxy[i] = lfxy; fxz[i] = lfxz;
 
             double gxa = gupxx[i]*Gamxxx[i] + gupyy[i]*Gamxyy[i] + gupzz[i]*Gamxzz[i]
                     + TWO * ( gupxy[i]*Gamxxy[i] + gupxz[i]*Gamxxz[i] + gupyz[i]*Gamxyz[i] );
@@ -763,74 +686,69 @@ int f_compute_rhs_bssn(int *ex, double &T,
                     + Gamxyz[i] * gzzx[i] + Gamyyz[i] * gzzy[i] + Gamzyz[i] * gzzz[i]
                 );
         }
-        RHS_KERNEL_TIMER_ADD(KB_RICCI_METRIC, timer_ricci_metric);
 
-        RHS_KERNEL_TIMER_DECL(timer_chi_lapse);
         // 22.3ms //
             fdderivs(ex,chi,fxx,fxy,fxz,fyy,fyz,fzz,X,Y,Z,SYM,SYM,SYM,Symmetry,Lev);
 
         // 7ms //
         for (int i=0;i<all;i+=1) {
-            const double inv_chin1 = ONE / chin1[i];
-            const double half_inv_chin1 = HALF * inv_chin1;
-            const double scaled_inv = F3o2 * inv_chin1;
-            const double cxx = fxx[i] - Gamxxx[i] * chix[i] - Gamyxx[i] * chiy[i] - Gamzxx[i] * chiz[i];
-            const double cxy = fxy[i] - Gamxxy[i] * chix[i] - Gamyxy[i] * chiy[i] - Gamzxy[i] * chiz[i];
-            const double cxz = fxz[i] - Gamxxz[i] * chix[i] - Gamyxz[i] * chiy[i] - Gamzxz[i] * chiz[i];
-            const double cyy = fyy[i] - Gamxyy[i] * chix[i] - Gamyyy[i] * chiy[i] - Gamzyy[i] * chiz[i];
-            const double cyz = fyz[i] - Gamxyz[i] * chix[i] - Gamyyz[i] * chiy[i] - Gamzyz[i] * chiz[i];
-            const double czz = fzz[i] - Gamxzz[i] * chix[i] - Gamyzz[i] * chiy[i] - Gamzzz[i] * chiz[i];
-            const double ricci_chi =
-                gupxx[i] * (cxx - scaled_inv * chix[i] * chix[i])
-                + gupyy[i] * (cyy - scaled_inv * chiy[i] * chiy[i])
-                + gupzz[i] * (czz - scaled_inv * chiz[i] * chiz[i])
-                + TWO * gupxy[i] * (cxy - scaled_inv * chix[i] * chiy[i])
-                + TWO * gupxz[i] * (cxz - scaled_inv * chix[i] * chiz[i])
-                + TWO * gupyz[i] * (cyz - scaled_inv * chiy[i] * chiz[i]);
-            f[i] = ricci_chi;
-            Rxx[i] = Rxx[i] + ( cxx - half_inv_chin1 * chix[i] * chix[i] + (dxx[i] + ONE) * ricci_chi ) * half_inv_chin1;
-            Ryy[i] = Ryy[i] + ( cyy - half_inv_chin1 * chiy[i] * chiy[i] + (dyy[i] + ONE) * ricci_chi ) * half_inv_chin1;
-            Rzz[i] = Rzz[i] + ( czz - half_inv_chin1 * chiz[i] * chiz[i] + (dzz[i] + ONE) * ricci_chi ) * half_inv_chin1;
+            fxx[i] = fxx[i] - Gamxxx[i] * chix[i] - Gamyxx[i] * chiy[i] - Gamzxx[i] * chiz[i];
+            fxy[i] = fxy[i] - Gamxxy[i] * chix[i] - Gamyxy[i] * chiy[i] - Gamzxy[i] * chiz[i];
+            fxz[i] = fxz[i] - Gamxxz[i] * chix[i] - Gamyxz[i] * chiy[i] - Gamzxz[i] * chiz[i];
+            fyy[i] = fyy[i] - Gamxyy[i] * chix[i] - Gamyyy[i] * chiy[i] - Gamzyy[i] * chiz[i];
+            fyz[i] = fyz[i] - Gamxyz[i] * chix[i] - Gamyyz[i] * chiy[i] - Gamzyz[i] * chiz[i];
+            fzz[i] = fzz[i] - Gamxzz[i] * chix[i] - Gamyzz[i] * chiy[i] - Gamzzz[i] * chiz[i];
+            f[i] =
+                gupxx[i] * (fxx[i] - (F3o2 / chin1[i]) * chix[i] * chix[i])
+                + gupyy[i] * (fyy[i] - (F3o2 / chin1[i]) * chiy[i] * chiy[i])
+                + gupzz[i] * (fzz[i] - (F3o2 / chin1[i]) * chiz[i] * chiz[i])
+                + TWO * gupxy[i] * (fxy[i] - (F3o2 / chin1[i]) * chix[i] * chiy[i])
+                + TWO * gupxz[i] * (fxz[i] - (F3o2 / chin1[i]) * chix[i] * chiz[i])
+                + TWO * gupyz[i] * (fyz[i] - (F3o2 / chin1[i]) * chiy[i] * chiz[i]);
+            Rxx[i] = Rxx[i] + ( fxx[i] - (chix[i] * chix[i]) / (chin1[i] * TWO) + (dxx[i] + ONE) * f[i] ) / (chin1[i] * TWO);
+            Ryy[i] = Ryy[i] + ( fyy[i] - (chiy[i] * chiy[i]) / (chin1[i] * TWO) + (dyy[i] + ONE) * f[i] ) / (chin1[i] * TWO);
+            Rzz[i] = Rzz[i] + ( fzz[i] - (chiz[i] * chiz[i]) / (chin1[i] * TWO) + (dzz[i] + ONE) * f[i] ) / (chin1[i] * TWO);
 
-            Rxy[i] = Rxy[i] + ( cxy - half_inv_chin1 * chix[i] * chiy[i] + gxy[i] * ricci_chi ) * half_inv_chin1;
-            Rxz[i] = Rxz[i] + ( cxz - half_inv_chin1 * chix[i] * chiz[i] + gxz[i] * ricci_chi ) * half_inv_chin1;
-            Ryz[i] = Ryz[i] + ( cyz - half_inv_chin1 * chiy[i] * chiz[i] + gyz[i] * ricci_chi ) * half_inv_chin1;
+            Rxy[i] = Rxy[i] + ( fxy[i] - (chix[i] * chiy[i]) / (chin1[i] * TWO) + gxy[i] * f[i] ) / (chin1[i] * TWO);
+            Rxz[i] = Rxz[i] + ( fxz[i] - (chix[i] * chiz[i]) / (chin1[i] * TWO) + gxz[i] * f[i] ) / (chin1[i] * TWO);
+            Ryz[i] = Ryz[i] + ( fyz[i] - (chiy[i] * chiz[i]) / (chin1[i] * TWO) + gyz[i] * f[i] ) / (chin1[i] * TWO);
         }
 
         // 24ms //
         fdderivs(ex,Lap,fxx,fxy,fxz,fyy,fyz,fzz,X,Y,Z,SYM,SYM,SYM,Symmetry,Lev);
+        fderivs(ex,chi,dtSfx_rhs,dtSfy_rhs,dtSfz_rhs,X,Y,Z,SYM,SYM,SYM,Symmetry,Lev);
 
         // 6ms //
         for (int i=0;i<all;i+=1) {
-            const double inv_chin1 = ONE / chin1[i];
-            const double gchi_x = (gupxx[i] * chix[i] + gupxy[i] * chiy[i] + gupxz[i] * chiz[i]) * inv_chin1;
-            const double gchi_y = (gupxy[i] * chix[i] + gupyy[i] * chiy[i] + gupyz[i] * chiz[i]) * inv_chin1;
-            const double gchi_z = (gupxz[i] * chix[i] + gupyz[i] * chiy[i] + gupzz[i] * chiz[i]) * inv_chin1;
+            /* gxxx,gxxy,gxxz (这里是“升指标后的chi导数/chi”那类量，你沿用原变量名即可) */
+            gxxx[i] = (gupxx[i] * chix[i] + gupxy[i] * chiy[i] + gupxz[i] * chiz[i]) / chin1[i];
+            gxxy[i] = (gupxy[i] * chix[i] + gupyy[i] * chiy[i] + gupyz[i] * chiz[i]) / chin1[i];
+            gxxz[i] = (gupxz[i] * chix[i] + gupyz[i] * chiy[i] + gupzz[i] * chiz[i]) / chin1[i];
 
             /* Christoffel 修正项 */
-            Gamxxx[i] = Gamxxx[i] - ( ((chix[i] + chix[i]) * inv_chin1) - (dxx[i] + ONE) * gchi_x ) * HALF;
-            Gamyxx[i] = Gamyxx[i] - ( 0.0           - (dxx[i] + ONE) * gchi_y ) * HALF;  /* 原式只有 -gxx*gxxy */
-            Gamzxx[i] = Gamzxx[i] - ( 0.0           - (dxx[i] + ONE) * gchi_z ) * HALF;
+            Gamxxx[i] = Gamxxx[i] - ( ((chix[i] + chix[i]) / chin1[i]) - (dxx[i] + ONE) * gxxx[i] ) * HALF;
+            Gamyxx[i] = Gamyxx[i] - ( 0.0           - (dxx[i] + ONE) * gxxy[i] ) * HALF;  /* 原式只有 -gxx*gxxy */
+            Gamzxx[i] = Gamzxx[i] - ( 0.0           - (dxx[i] + ONE) * gxxz[i] ) * HALF;
 
-            Gamxyy[i] = Gamxyy[i] - ( 0.0           - (dyy[i] + ONE) * gchi_x ) * HALF;
-            Gamyyy[i] = Gamyyy[i] - ( ((chiy[i] + chiy[i]) * inv_chin1) - (dyy[i] + ONE) * gchi_y ) * HALF;
-            Gamzyy[i] = Gamzyy[i] - ( 0.0           - (dyy[i] + ONE) * gchi_z ) * HALF;
+            Gamxyy[i] = Gamxyy[i] - ( 0.0           - (dyy[i] + ONE) * gxxx[i] ) * HALF;
+            Gamyyy[i] = Gamyyy[i] - ( ((chiy[i] + chiy[i]) / chin1[i]) - (dyy[i] + ONE) * gxxy[i] ) * HALF;
+            Gamzyy[i] = Gamzyy[i] - ( 0.0           - (dyy[i] + ONE) * gxxz[i] ) * HALF;
 
-            Gamxzz[i] = Gamxzz[i] - ( 0.0           - (dzz[i] + ONE) * gchi_x ) * HALF;
-            Gamyzz[i] = Gamyzz[i] - ( 0.0           - (dzz[i] + ONE) * gchi_y ) * HALF;
-            Gamzzz[i] = Gamzzz[i] - ( ((chiz[i] + chiz[i]) * inv_chin1) - (dzz[i] + ONE) * gchi_z ) * HALF;
+            Gamxzz[i] = Gamxzz[i] - ( 0.0           - (dzz[i] + ONE) * gxxx[i] ) * HALF;
+            Gamyzz[i] = Gamyzz[i] - ( 0.0           - (dzz[i] + ONE) * gxxy[i] ) * HALF;
+            Gamzzz[i] = Gamzzz[i] - ( ((chiz[i] + chiz[i]) / chin1[i]) - (dzz[i] + ONE) * gxxz[i] ) * HALF;
 
-            Gamxxy[i] = Gamxxy[i] - ( ( chiy[i] * inv_chin1)      - gxy[i] * gchi_x ) * HALF;
-            Gamyxy[i] = Gamyxy[i] - ( ( chix[i] * inv_chin1)      - gxy[i] * gchi_y ) * HALF;
-            Gamzxy[i] = Gamzxy[i] - ( 0.0           - gxy[i] * gchi_z ) * HALF;
+            Gamxxy[i] = Gamxxy[i] - ( ( chiy[i] / chin1[i])      - gxy[i] * gxxx[i] ) * HALF;
+            Gamyxy[i] = Gamyxy[i] - ( ( chix[i] / chin1[i])      - gxy[i] * gxxy[i] ) * HALF;
+            Gamzxy[i] = Gamzxy[i] - ( 0.0           - gxy[i] * gxxz[i] ) * HALF;
 
-            Gamxxz[i] = Gamxxz[i] - ( ( chiz[i] * inv_chin1)      - gxz[i] * gchi_x ) * HALF;
-            Gamyxz[i] = Gamyxz[i] - ( 0.0           - gxz[i] * gchi_y ) * HALF;
-            Gamzxz[i] = Gamzxz[i] - ( ( chix[i] * inv_chin1)      - gxz[i] * gchi_z ) * HALF;
+            Gamxxz[i] = Gamxxz[i] - ( ( chiz[i] / chin1[i])      - gxz[i] * gxxx[i] ) * HALF;
+            Gamyxz[i] = Gamyxz[i] - ( 0.0           - gxz[i] * gxxy[i] ) * HALF;
+            Gamzxz[i] = Gamzxz[i] - ( ( chix[i] / chin1[i])      - gxz[i] * gxxz[i] ) * HALF;
 
-            Gamxyz[i] = Gamxyz[i] - ( 0.0           - gyz[i] * gchi_x ) * HALF;
-            Gamyyz[i] = Gamyyz[i] - ( ( chiz[i] * inv_chin1)      - gyz[i] * gchi_y ) * HALF;
-            Gamzyz[i] = Gamzyz[i] - ( ( chiy[i] * inv_chin1)      - gyz[i] * gchi_z ) * HALF;
+            Gamxyz[i] = Gamxyz[i] - ( 0.0           - gyz[i] * gxxx[i] ) * HALF;
+            Gamyyz[i] = Gamyyz[i] - ( ( chiz[i] / chin1[i])      - gyz[i] * gxxy[i] ) * HALF;
+            Gamzyz[i] = Gamzyz[i] - ( ( chiy[i] / chin1[i])      - gyz[i] * gxxz[i] ) * HALF;
 
             /* fxx..fyz 修正：减去 Γ * ∂Lap */
             fxx[i] = fxx[i] - Gamxxx[i] * Lapx[i] - Gamyxx[i] * Lapy[i] - Gamzxx[i] * Lapz[i];
@@ -844,8 +762,6 @@ int f_compute_rhs_bssn(int *ex, double &T,
             trK_rhs[i] =  gupxx[i] * fxx[i] + gupyy[i] * fyy[i] + gupzz[i] * fzz[i]
                     +  TWO * ( gupxy[i] * fxy[i] + gupxz[i] * fxz[i] + gupyz[i] * fyz[i] );
         }
-        RHS_KERNEL_TIMER_ADD(KB_CHI_LAPSE, timer_chi_lapse);
-        RHS_KERNEL_TIMER_DECL(timer_aij_trk_gauge);
         // 2.5ms //
         for (int i=0;i<all;i+=1) {
             const double divb = betaxx[i] + betayy[i] + betazz[i];
@@ -1106,9 +1022,16 @@ int f_compute_rhs_bssn(int *ex, double &T,
                         + gupyz[i] * dtSfy_rhs[i] * dtSfz_rhs[i] );
 
             #if (GAUGE == 2)
-            reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / pow( (ONE - sqrt(chin1[i])), 2.0 );
+            {
+                const double chi_sqrt = sqrt(chin1[i]);
+                const double damping = ONE - chi_sqrt;
+                reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / (damping * damping);
+            }
             #else
-            reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / pow( (ONE - chin1[i]), 2.0 );
+            {
+                const double damping = ONE - chin1[i];
+                reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / (damping * damping);
+            }
             #endif
 
             dtSfx_rhs[i] = Gamx_rhs[i] - reta[i] * dtSfx[i];
@@ -1124,9 +1047,16 @@ int f_compute_rhs_bssn(int *ex, double &T,
                         + gupyz[i] * dtSfy_rhs[i] * dtSfz_rhs[i] );
 
             #if (GAUGE == 4)
-            reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / pow( (ONE - sqrt(chin1[i])), 2.0 );
+            {
+                const double chi_sqrt = sqrt(chin1[i]);
+                const double damping = ONE - chi_sqrt;
+                reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / (damping * damping);
+            }
             #else
-            reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / pow( (ONE - chin1[i]), 2.0 );
+            {
+                const double damping = ONE - chin1[i];
+                reta[i] = 1.31 / 2.0 * sqrt( reta[i] / chin1[i] ) / (damping * damping);
+            }
             #endif
 
             betax_rhs[i] = FF * Gamx[i] - reta[i] * betax[i];
@@ -1146,8 +1076,6 @@ int f_compute_rhs_bssn(int *ex, double &T,
             dtSfz_rhs[i] = Gamz_rhs[i] - reta[i] * dtSfz[i];
             #endif
         }
-        RHS_KERNEL_TIMER_ADD(KB_AIJ_TRK_GAUGE, timer_aij_trk_gauge);
-        RHS_KERNEL_TIMER_DECL(timer_ko_constraint);
         // advection + KO dissipation with shared symmetry buffer
         lopsided_kodis(ex,X,Y,Z,dxx,gxx_rhs,betax,betay,betaz,Symmetry,SSS,eps);
         lopsided_kodis(ex,X,Y,Z,Gamz,Gamz_rhs,betax,betay,betaz,Symmetry,SSA,eps);
@@ -1279,7 +1207,6 @@ int f_compute_rhs_bssn(int *ex, double &T,
                 movz_Res[i] = movz_Res[i] - F2o3*Kz[i] - F8*PI*Sz[i];
             }
         }
-        RHS_KERNEL_TIMER_ADD(KB_KO_CONSTRAINT, timer_ko_constraint);
 
 
             

AMSS_NCKU_source/cgh.C

@@ -23,10 +23,14 @@ using namespace std;
 #include <mpi.h>
 
 #include "macrodef.h"
-#include "misc.h"
-#include "cgh.h"
-#include "Parallel.h"
-#include "parameters.h"
+#include "misc.h"
+#include "cgh.h"
+#include "Parallel.h"
+#include "parameters.h"
+#ifdef USE_GPU
+#include "bssn_gpu.h"
+#include "bssn_cuda_ops.h"
+#endif
 
 //================================================================================================
 
@@ -881,13 +885,20 @@ void cgh::recompose_cgh(int nprocs, bool *lev_flag,
       tmPat = construct_patchlist(lev, Symmetry);
       // tmPat construction completes
       Parallel::distribute(tmPat, nprocs, ingfs, fngfs, false);
-      //    checkPatchList(tmPat,true);
-      bool CC = (lev > trfls);
-      Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
-
-      Parallel::KillBlocks(PatL[lev]);
-      PatL[lev]->destroyList();
-      PatL[lev] = tmPat;
+      //    checkPatchList(tmPat,true);
+      bool CC = (lev > trfls);
+      Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
+
+#ifdef USE_GPU
+      bssn_gpu_clear_cached_device_buffers();
+      bssn_gpu_release_pinned_host_buffers();
+      bssn_cuda_release_rk4_caches();
+      bssn_cuda_release_interp_caches();
+      patch_release_interp_plan_cache();
+#endif
+      Parallel::KillBlocks(PatL[lev]);
+      PatL[lev]->destroyList();
+      PatL[lev] = tmPat;
 #if (RPB == 1)
       Parallel::destroypsuList_bam(bdsul[lev]);
       Parallel::destroypsuList_bam(rsul[lev]);
@@ -910,13 +921,20 @@ void cgh::recompose_cgh(int nprocs, bool *lev_flag,
       tmPat = construct_patchlist(lev, Symmetry);
       // tmPat construction completes
       Parallel::distribute(tmPat, end_rank[lev] - start_rank[lev] + 1, ingfs, fngfs, false, start_rank[lev], end_rank[lev]);
-      //    checkPatchList(tmPat,true);
-      bool CC = (lev > trfls);
-      Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
-
-      Parallel::KillBlocks(PatL[lev]);
-      PatL[lev]->destroyList();
-      PatL[lev] = tmPat;
+      //    checkPatchList(tmPat,true);
+      bool CC = (lev > trfls);
+      Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
+
+#ifdef USE_GPU
+      bssn_gpu_clear_cached_device_buffers();
+      bssn_gpu_release_pinned_host_buffers();
+      bssn_cuda_release_rk4_caches();
+      bssn_cuda_release_interp_caches();
+      patch_release_interp_plan_cache();
+#endif
+      Parallel::KillBlocks(PatL[lev]);
+      PatL[lev]->destroyList();
+      PatL[lev] = tmPat;
 #if (RPB == 1)
 #error "not support yet"
 #endif
@@ -1518,13 +1536,20 @@ void cgh::recompose_cgh_Onelevel(int nprocs, int lev,
   tmPat = construct_patchlist(lev, Symmetry);
   // tmPat construction completes
   Parallel::distribute(tmPat, nprocs, ingfs, fngfs, false);
-  //    checkPatchList(tmPat,true);
-  bool CC = (lev > trfls);
-  Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
-
-  Parallel::KillBlocks(PatL[lev]);
-  PatL[lev]->destroyList();
-  PatL[lev] = tmPat;
+  //    checkPatchList(tmPat,true);
+  bool CC = (lev > trfls);
+  Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
+
+#ifdef USE_GPU
+  bssn_gpu_clear_cached_device_buffers();
+  bssn_gpu_release_pinned_host_buffers();
+  bssn_cuda_release_rk4_caches();
+  bssn_cuda_release_interp_caches();
+  patch_release_interp_plan_cache();
+#endif
+  Parallel::KillBlocks(PatL[lev]);
+  PatL[lev]->destroyList();
+  PatL[lev] = tmPat;
 }
 #elif (PSTR == 1 || PSTR == 2 || PSTR == 3)
 #warning "recompose_cgh_Onelevel is not implimented yet"
@@ -1540,14 +1565,21 @@ void cgh::recompose_cgh_Onelevel(int nprocs, int lev,
   // tmPat construction completes
   Parallel::distribute(tmPat, end_rank[lev] - start_rank[lev] + 1, ingfs, fngfs, false, start_rank[lev], end_rank[lev]);
   misc::tillherecheck(Commlev[lev], start_rank[lev], "after distribute");
-  //    checkPatchList(tmPat,true);
-  bool CC = (lev > trfls);
-  Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
-  misc::tillherecheck(Commlev[lev], start_rank[lev], "after fill_level_data");
-
-  Parallel::KillBlocks(PatL[lev]);
-  PatL[lev]->destroyList();
-  PatL[lev] = tmPat;
+  //    checkPatchList(tmPat,true);
+  bool CC = (lev > trfls);
+  Parallel::fill_level_data(tmPat, PatL[lev], PatL[lev - 1], OldList, StateList, FutureList, tmList, Symmetry, BB, CC);
+  misc::tillherecheck(Commlev[lev], start_rank[lev], "after fill_level_data");
+
+#ifdef USE_GPU
+  bssn_gpu_clear_cached_device_buffers();
+  bssn_gpu_release_pinned_host_buffers();
+  bssn_cuda_release_rk4_caches();
+  bssn_cuda_release_interp_caches();
+  patch_release_interp_plan_cache();
+#endif
+  Parallel::KillBlocks(PatL[lev]);
+  PatL[lev]->destroyList();
+  PatL[lev] = tmPat;
 }
 
 

AMSS_NCKU_source/fdderivs_sh_c.C

@@ -1,321 +0,0 @@
-#include "macrodef.h"
-#include "share_func.h"
-
-/*
- * fdderivs_sh — second derivatives on shell patch in (rho, sigma, R) coords.
- * Same stencil coefficients as Cartesian fdderivs. Uses symmetry_stbd.
- */
-extern "C" void fdderivs_sh_(const int ex[3],
-                 const double *f,
-                 double *fxx, double *fxy, double *fxz,
-                 double *fyy, double *fyz, double *fzz,
-                 const double *X, const double *Y, const double *Z,
-                 double SYM1, double SYM2, double SYM3,
-                 int Symmetry, int onoff, int sst)
-{
-    (void)SYM3; (void)onoff; (void)sst;
-
-    const int NO_SYMM=0, EQ_SYMM=1, OCTANT=2;
-    const double ZEO=0.0, ONE=1.0, TWO=2.0, F1o4=2.5e-1;
-    const double F8=8.0, F16=16.0, F30=30.0, F1o12=ONE/12.0, F1o144=ONE/144.0;
-    const double F9=9.0, F45=45.0, F60=60.0, F27=27.0, F270=270.0, F490=490.0;
-    const double F1o180=ONE/180.0, F1o3600=ONE/3600.0;
-    const double F32=32.0, F128=128.0, F168=168.0, F672=672.0, F840=840.0;
-    const double F1008=1008.0, F8064=8064.0, F14350=14350.0;
-    const double F1o5040=ONE/5040.0, F1o705600=ONE/705600.0;
-
-    const int ex1=ex[0], ex2=ex[1], ex3=ex[2];
-    const double dX=X[1]-X[0], dY=Y[1]-Y[0], dZ=Z[1]-Z[0];
-    const int imaxF=ex1, jmaxF=ex2, kmaxF=ex3;
-    const double SoA[2]={SYM1,SYM2};
-
-#if (ghost_width == 2)
-    {
-        const int ord=1;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=0;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=0;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=0;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        static double *fh_buf=NULL;static size_t cap=0;
-        if(fh_size>cap){free(fh_buf);fh_buf=(double*)aligned_alloc(64,fh_size*sizeof(double));cap=fh_size;}
-        double *fh=fh_buf;if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const double Sdxdx=ONE/(dX*dX),Sdydy=ONE/(dY*dY),Sdzdz=ONE/(dZ*dZ);
-        const double Sdxdy=F1o4/(dX*dY),Sdxdz=F1o4/(dX*dZ),Sdydz=F1o4/(dY*dZ);
-        const size_t all=(size_t)ex1*ex2*ex3;
-        for(size_t p=0;p<all;++p){fxx[p]=fyy[p]=fzz[p]=ZEO;fxy[p]=fxz[p]=fyz[p]=ZEO;}
-
-        const int i2_lo=(iminF>0)?iminF:0,j2_lo=(jminF>0)?jminF:0,k2_lo=1,i2_hi=ex1-2,j2_hi=ex2-2,k2_hi=ex3-2;
-        #define FH(iF,jF,kF) fh[idx_fh_stbd(iF,jF,kF,ord,ex)]
-        if(i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi){
-            for(int k0=k2_lo;k0<=k2_hi;++k0){const int kF=k0+1;
-            for(int j0=j2_lo;j0<=j2_hi;++j0){const int jF=j0+1;
-            for(int i0=i2_lo;i0<=i2_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-                fxx[p]=Sdxdx*(FH(iF-1,jF,kF)-TWO*FH(iF,jF,kF)+FH(iF+1,jF,kF));
-                fyy[p]=Sdydy*(FH(iF,jF-1,kF)-TWO*FH(iF,jF,kF)+FH(iF,jF+1,kF));
-                fzz[p]=Sdzdz*(FH(iF,jF,kF-1)-TWO*FH(iF,jF,kF)+FH(iF,jF,kF+1));
-                fxy[p]=Sdxdy*(FH(iF-1,jF-1,kF)-FH(iF+1,jF-1,kF)-FH(iF-1,jF+1,kF)+FH(iF+1,jF+1,kF));
-                fxz[p]=Sdxdz*(FH(iF-1,jF,kF-1)-FH(iF+1,jF,kF-1)-FH(iF-1,jF,kF+1)+FH(iF+1,jF,kF+1));
-                fyz[p]=Sdydz*(FH(iF,jF-1,kF-1)-FH(iF,jF+1,kF-1)-FH(iF,jF-1,kF+1)+FH(iF,jF+1,kF+1));
-            }}}
-        }
-        #undef FH
-        return;
-    }
-#elif (ghost_width == 3)
-    {
-        const int ord=2;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-1;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-1;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-1;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        static double *fh_buf=NULL;static size_t cap=0;
-        if(fh_size>cap){free(fh_buf);fh_buf=(double*)aligned_alloc(64,fh_size*sizeof(double));cap=fh_size;}
-        double *fh=fh_buf;if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const double Sdxdx=ONE/(dX*dX),Sdydy=ONE/(dY*dY),Sdzdz=ONE/(dZ*dZ);
-        const double Fdxdx=F1o12/(dX*dX),Fdydy=F1o12/(dY*dY),Fdzdz=F1o12/(dZ*dZ);
-        const double Sdxdy=F1o4/(dX*dY),Sdxdz=F1o4/(dX*dZ),Sdydz=F1o4/(dY*dZ);
-        const double Fdxdy=F1o144/(dX*dY),Fdxdz=F1o144/(dX*dZ),Fdydz=F1o144/(dY*dZ);
-        const size_t all=(size_t)ex1*ex2*ex3;
-        for(size_t p=0;p<all;++p){fxx[p]=fyy[p]=fzz[p]=fxy[p]=fxz[p]=fyz[p]=ZEO;}
-
-        const int i2_lo=(iminF>0)?iminF:0,j2_lo=(jminF>0)?jminF:0,k2_lo=1,i2_hi=ex1-2,j2_hi=ex2-2,k2_hi=ex3-2;
-        const int i4_lo=(iminF+1>0)?iminF+1:0,j4_lo=(jminF+1>0)?jminF+1:0,k4_lo=2,i4_hi=ex1-3,j4_hi=ex2-3,k4_hi=ex3-3;
-        const int has4=(i4_lo<=i4_hi&&j4_lo<=j4_hi&&k4_lo<=k4_hi);
-        #define FH(iF,jF,kF) fh[idx_fh_stbd(iF,jF,kF,ord,ex)]
-
-        if(i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi){
-            for(int k0=k2_lo;k0<=k2_hi;++k0){const int kF=k0+1;
-            for(int j0=j2_lo;j0<=j2_hi;++j0){const int jF=j0+1;
-            for(int i0=i2_lo;i0<=i2_hi;++i0){
-                if(has4&&i0>=i4_lo&&i0<=i4_hi&&j0>=j4_lo&&j0<=j4_hi&&k0>=k4_lo&&k0<=k4_hi)continue;
-                const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-                fxx[p]=Sdxdx*(FH(iF-1,jF,kF)-TWO*FH(iF,jF,kF)+FH(iF+1,jF,kF));
-                fyy[p]=Sdydy*(FH(iF,jF-1,kF)-TWO*FH(iF,jF,kF)+FH(iF,jF+1,kF));
-                fzz[p]=Sdzdz*(FH(iF,jF,kF-1)-TWO*FH(iF,jF,kF)+FH(iF,jF,kF+1));
-                fxy[p]=Sdxdy*(FH(iF-1,jF-1,kF)-FH(iF+1,jF-1,kF)-FH(iF-1,jF+1,kF)+FH(iF+1,jF+1,kF));
-                fxz[p]=Sdxdz*(FH(iF-1,jF,kF-1)-FH(iF+1,jF,kF-1)-FH(iF-1,jF,kF+1)+FH(iF+1,jF,kF+1));
-                fyz[p]=Sdydz*(FH(iF,jF-1,kF-1)-FH(iF,jF+1,kF-1)-FH(iF,jF-1,kF+1)+FH(iF,jF+1,kF+1));
-            }}}
-        }
-        if(has4){
-            for(int k0=k4_lo;k0<=k4_hi;++k0){const int kF=k0+1;
-            for(int j0=j4_lo;j0<=j4_hi;++j0){const int jF=j0+1;
-            for(int i0=i4_lo;i0<=i4_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-                fxx[p]=Fdxdx*(-FH(iF-2,jF,kF)+F16*FH(iF-1,jF,kF)-F30*FH(iF,jF,kF)-FH(iF+2,jF,kF)+F16*FH(iF+1,jF,kF));
-                fyy[p]=Fdydy*(-FH(iF,jF-2,kF)+F16*FH(iF,jF-1,kF)-F30*FH(iF,jF,kF)-FH(iF,jF+2,kF)+F16*FH(iF,jF+1,kF));
-                fzz[p]=Fdzdz*(-FH(iF,jF,kF-2)+F16*FH(iF,jF,kF-1)-F30*FH(iF,jF,kF)-FH(iF,jF,kF+2)+F16*FH(iF,jF,kF+1));
-                {const double t_jm2=(FH(iF-2,jF-2,kF)-F8*FH(iF-1,jF-2,kF)+F8*FH(iF+1,jF-2,kF)-FH(iF+2,jF-2,kF));
-                 const double t_jm1=(FH(iF-2,jF-1,kF)-F8*FH(iF-1,jF-1,kF)+F8*FH(iF+1,jF-1,kF)-FH(iF+2,jF-1,kF));
-                 const double t_jp1=(FH(iF-2,jF+1,kF)-F8*FH(iF-1,jF+1,kF)+F8*FH(iF+1,jF+1,kF)-FH(iF+2,jF+1,kF));
-                 const double t_jp2=(FH(iF-2,jF+2,kF)-F8*FH(iF-1,jF+2,kF)+F8*FH(iF+1,jF+2,kF)-FH(iF+2,jF+2,kF));
-                 fxy[p]=Fdxdy*(t_jm2-F8*t_jm1+F8*t_jp1-t_jp2);}
-                {const double t_km2=(FH(iF-2,jF,kF-2)-F8*FH(iF-1,jF,kF-2)+F8*FH(iF+1,jF,kF-2)-FH(iF+2,jF,kF-2));
-                 const double t_km1=(FH(iF-2,jF,kF-1)-F8*FH(iF-1,jF,kF-1)+F8*FH(iF+1,jF,kF-1)-FH(iF+2,jF,kF-1));
-                 const double t_kp1=(FH(iF-2,jF,kF+1)-F8*FH(iF-1,jF,kF+1)+F8*FH(iF+1,jF,kF+1)-FH(iF+2,jF,kF+1));
-                 const double t_kp2=(FH(iF-2,jF,kF+2)-F8*FH(iF-1,jF,kF+2)+F8*FH(iF+1,jF,kF+2)-FH(iF+2,jF,kF+2));
-                 fxz[p]=Fdxdz*(t_km2-F8*t_km1+F8*t_kp1-t_kp2);}
-                {const double t_km2=(FH(iF,jF-2,kF-2)-F8*FH(iF,jF-1,kF-2)+F8*FH(iF,jF+1,kF-2)-FH(iF,jF+2,kF-2));
-                 const double t_km1=(FH(iF,jF-2,kF-1)-F8*FH(iF,jF-1,kF-1)+F8*FH(iF,jF+1,kF-1)-FH(iF,jF+2,kF-1));
-                 const double t_kp1=(FH(iF,jF-2,kF+1)-F8*FH(iF,jF-1,kF+1)+F8*FH(iF,jF+1,kF+1)-FH(iF,jF+2,kF+1));
-                 const double t_kp2=(FH(iF,jF-2,kF+2)-F8*FH(iF,jF-1,kF+2)+F8*FH(iF,jF+1,kF+2)-FH(iF,jF+2,kF+2));
-                 fyz[p]=Fdydz*(t_km2-F8*t_km1+F8*t_kp1-t_kp2);}
-            }}}
-        }
-        #undef FH
-        return;
-    }
-#elif (ghost_width == 4)
-    {
-        const int ord=3;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-2;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-2;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-2;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        static double *fh_buf=NULL;static size_t cap=0;
-        if(fh_size>cap){free(fh_buf);fh_buf=(double*)aligned_alloc(64,fh_size*sizeof(double));cap=fh_size;}
-        double *fh=fh_buf;if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const double Sdxdx=ONE/(dX*dX),Sdydy=ONE/(dY*dY),Sdzdz=ONE/(dZ*dZ);
-        const double Fdxdx=F1o12/(dX*dX),Fdydy=F1o12/(dY*dY),Fdzdz=F1o12/(dZ*dZ);
-        const double Xdxdx=F1o180/(dX*dX),Xdydy=F1o180/(dY*dY),Xdzdz=F1o180/(dZ*dZ);
-        const double Sdxdy=F1o4/(dX*dY),Sdxdz=F1o4/(dX*dZ),Sdydz=F1o4/(dY*dZ);
-        const double Fdxdy=F1o144/(dX*dY),Fdxdz=F1o144/(dX*dZ),Fdydz=F1o144/(dY*dZ);
-        const double Xdxdy=F1o3600/(dX*dY),Xdxdz=F1o3600/(dX*dZ),Xdydz=F1o3600/(dY*dZ);
-        const size_t all=(size_t)ex1*ex2*ex3;
-        for(size_t p=0;p<all;++p){fxx[p]=fyy[p]=fzz[p]=fxy[p]=fxz[p]=fyz[p]=ZEO;}
-
-        const int i2_lo=(iminF>0)?iminF:0,j2_lo=(jminF>0)?jminF:0,k2_lo=1,i2_hi=ex1-2,j2_hi=ex2-2,k2_hi=ex3-2;
-        const int i4_lo=(iminF+1>0)?iminF+1:0,j4_lo=(jminF+1>0)?jminF+1:0,k4_lo=2,i4_hi=ex1-3,j4_hi=ex2-3,k4_hi=ex3-3;
-        const int i6_lo=(iminF+2>0)?iminF+2:0,j6_lo=(jminF+2>0)?jminF+2:0,k6_lo=3,i6_hi=ex1-4,j6_hi=ex2-4,k6_hi=ex3-4;
-        const int has4=(i4_lo<=i4_hi&&j4_lo<=j4_hi&&k4_lo<=k4_hi),has6=(i6_lo<=i6_hi&&j6_lo<=j6_hi&&k6_lo<=k6_hi);
-        #define FH(iF,jF,kF) fh[idx_fh_stbd(iF,jF,kF,ord,ex)]
-
-        if(i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi){for(int k0=k2_lo;k0<=k2_hi;++k0){const int kF=k0+1;
-        for(int j0=j2_lo;j0<=j2_hi;++j0){const int jF=j0+1;
-        for(int i0=i2_lo;i0<=i2_hi;++i0){_Bool in4=has4&&i0>=i4_lo&&i0<=i4_hi&&j0>=j4_lo&&j0<=j4_hi&&k0>=k4_lo&&k0<=k4_hi;if(in4)continue;
-            const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fxx[p]=Sdxdx*(FH(iF-1,jF,kF)-TWO*FH(iF,jF,kF)+FH(iF+1,jF,kF));
-            fyy[p]=Sdydy*(FH(iF,jF-1,kF)-TWO*FH(iF,jF,kF)+FH(iF,jF+1,kF));
-            fzz[p]=Sdzdz*(FH(iF,jF,kF-1)-TWO*FH(iF,jF,kF)+FH(iF,jF,kF+1));
-            fxy[p]=Sdxdy*(FH(iF-1,jF-1,kF)-FH(iF+1,jF-1,kF)-FH(iF-1,jF+1,kF)+FH(iF+1,jF+1,kF));
-            fxz[p]=Sdxdz*(FH(iF-1,jF,kF-1)-FH(iF+1,jF,kF-1)-FH(iF-1,jF,kF+1)+FH(iF+1,jF,kF+1));
-            fyz[p]=Sdydz*(FH(iF,jF-1,kF-1)-FH(iF,jF+1,kF-1)-FH(iF,jF-1,kF+1)+FH(iF,jF+1,kF+1));
-        }}}}
-        if(has4){for(int k0=k4_lo;k0<=k4_hi;++k0){const int kF=k0+1;
-        for(int j0=j4_lo;j0<=j4_hi;++j0){const int jF=j0+1;
-        for(int i0=i4_lo;i0<=i4_hi;++i0){if(has6&&i0>=i6_lo&&i0<=i6_hi&&j0>=j6_lo&&j0<=j6_hi&&k0>=k6_lo&&k0<=k6_hi)continue;
-            const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fxx[p]=Fdxdx*(-FH(iF-2,jF,kF)+F16*FH(iF-1,jF,kF)-F30*FH(iF,jF,kF)-FH(iF+2,jF,kF)+F16*FH(iF+1,jF,kF));
-            fyy[p]=Fdydy*(-FH(iF,jF-2,kF)+F16*FH(iF,jF-1,kF)-F30*FH(iF,jF,kF)-FH(iF,jF+2,kF)+F16*FH(iF,jF+1,kF));
-            fzz[p]=Fdzdz*(-FH(iF,jF,kF-2)+F16*FH(iF,jF,kF-1)-F30*FH(iF,jF,kF)-FH(iF,jF,kF+2)+F16*FH(iF,jF,kF+1));
-            {const double t_jm2=(FH(iF-2,jF-2,kF)-F8*FH(iF-1,jF-2,kF)+F8*FH(iF+1,jF-2,kF)-FH(iF+2,jF-2,kF));
-             const double t_jm1=(FH(iF-2,jF-1,kF)-F8*FH(iF-1,jF-1,kF)+F8*FH(iF+1,jF-1,kF)-FH(iF+2,jF-1,kF));
-             const double t_jp1=(FH(iF-2,jF+1,kF)-F8*FH(iF-1,jF+1,kF)+F8*FH(iF+1,jF+1,kF)-FH(iF+2,jF+1,kF));
-             const double t_jp2=(FH(iF-2,jF+2,kF)-F8*FH(iF-1,jF+2,kF)+F8*FH(iF+1,jF+2,kF)-FH(iF+2,jF+2,kF));
-             fxy[p]=Fdxdy*(t_jm2-F8*t_jm1+F8*t_jp1-t_jp2);}
-            {const double t_km2=(FH(iF-2,jF,kF-2)-F8*FH(iF-1,jF,kF-2)+F8*FH(iF+1,jF,kF-2)-FH(iF+2,jF,kF-2));
-             const double t_km1=(FH(iF-2,jF,kF-1)-F8*FH(iF-1,jF,kF-1)+F8*FH(iF+1,jF,kF-1)-FH(iF+2,jF,kF-1));
-             const double t_kp1=(FH(iF-2,jF,kF+1)-F8*FH(iF-1,jF,kF+1)+F8*FH(iF+1,jF,kF+1)-FH(iF+2,jF,kF+1));
-             const double t_kp2=(FH(iF-2,jF,kF+2)-F8*FH(iF-1,jF,kF+2)+F8*FH(iF+1,jF,kF+2)-FH(iF+2,jF,kF+2));
-             fxz[p]=Fdxdz*(t_km2-F8*t_km1+F8*t_kp1-t_kp2);}
-            {const double t_km2=(FH(iF,jF-2,kF-2)-F8*FH(iF,jF-1,kF-2)+F8*FH(iF,jF+1,kF-2)-FH(iF,jF+2,kF-2));
-             const double t_km1=(FH(iF,jF-2,kF-1)-F8*FH(iF,jF-1,kF-1)+F8*FH(iF,jF+1,kF-1)-FH(iF,jF+2,kF-1));
-             const double t_kp1=(FH(iF,jF-2,kF+1)-F8*FH(iF,jF-1,kF+1)+F8*FH(iF,jF+1,kF+1)-FH(iF,jF+2,kF+1));
-             const double t_kp2=(FH(iF,jF-2,kF+2)-F8*FH(iF,jF-1,kF+2)+F8*FH(iF,jF+1,kF+2)-FH(iF,jF+2,kF+2));
-             fyz[p]=Fdydz*(t_km2-F8*t_km1+F8*t_kp1-t_kp2);}
-        }}}}
-        if(has6){for(int k0=k6_lo;k0<=k6_hi;++k0){const int kF=k0+1;
-        for(int j0=j6_lo;j0<=j6_hi;++j0){const int jF=j0+1;
-        for(int i0=i6_lo;i0<=i6_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fxx[p]=Xdxdx*(TWO*FH(iF-3,jF,kF)-F27*FH(iF-2,jF,kF)+F270*FH(iF-1,jF,kF)-F490*FH(iF,jF,kF)+F270*FH(iF+1,jF,kF)-F27*FH(iF+2,jF,kF)+TWO*FH(iF+3,jF,kF));
-            fyy[p]=Xdydy*(TWO*FH(iF,jF-3,kF)-F27*FH(iF,jF-2,kF)+F270*FH(iF,jF-1,kF)-F490*FH(iF,jF,kF)+F270*FH(iF,jF+1,kF)-F27*FH(iF,jF+2,kF)+TWO*FH(iF,jF+3,kF));
-            fzz[p]=Xdzdz*(TWO*FH(iF,jF,kF-3)-F27*FH(iF,jF,kF-2)+F270*FH(iF,jF,kF-1)-F490*FH(iF,jF,kF)+F270*FH(iF,jF,kF+1)-F27*FH(iF,jF,kF+2)+TWO*FH(iF,jF,kF+3));
-            #define XS6(JF,KFDUMMY) (-FH(iF-3,JF,KFDUMMY)+F9*FH(iF-2,JF,KFDUMMY)-F45*FH(iF-1,JF,KFDUMMY)+F45*FH(iF+1,JF,KFDUMMY)-F9*FH(iF+2,JF,KFDUMMY)+FH(iF+3,JF,KFDUMMY))
-            fxy[p]=Xdxdy*(-XS6(jF-3,kF)+F9*XS6(jF-2,kF)-F45*XS6(jF-1,kF)+F45*XS6(jF+1,kF)-F9*XS6(jF+2,kF)+XS6(jF+3,kF));
-            fxz[p]=Xdxdz*(-XS6(jF,kF-3)+F9*XS6(jF,kF-2)-F45*XS6(jF,kF-1)+F45*XS6(jF,kF+1)-F9*XS6(jF,kF+2)+XS6(jF,kF+3));
-            #undef XS6
-            #define YS6(JF,KFDUMMY) (-FH(iF,JF-3,KFDUMMY)+F9*FH(iF,JF-2,KFDUMMY)-F45*FH(iF,JF-1,KFDUMMY)+F45*FH(iF,JF+1,KFDUMMY)-F9*FH(iF,JF+2,KFDUMMY)+FH(iF,JF+3,KFDUMMY))
-            fyz[p]=Xdydz*(-YS6(jF,kF-3)+F9*YS6(jF,kF-2)-F45*YS6(jF,kF-1)+F45*YS6(jF,kF+1)-F9*YS6(jF,kF+2)+YS6(jF,kF+3));
-            #undef YS6
-        }}}}
-        #undef FH
-        return;
-    }
-#elif (ghost_width == 5)
-    {
-        /* 8th-order shell second derivatives — inherits 8th-order stencil coeffs from Cartesian */
-        const int ord=4;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-3;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-3;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-3;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        static double *fh_buf=NULL;static size_t cap=0;
-        if(fh_size>cap){free(fh_buf);fh_buf=(double*)aligned_alloc(64,fh_size*sizeof(double));cap=fh_size;}
-        double *fh=fh_buf;if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const double Sdxdx=ONE/(dX*dX),Sdydy=ONE/(dY*dY),Sdzdz=ONE/(dZ*dZ);
-        const double Fdxdx=F1o12/(dX*dX),Fdydy=F1o12/(dY*dY),Fdzdz=F1o12/(dZ*dZ);
-        const double Xdxdx=F1o180/(dX*dX),Xdydy=F1o180/(dY*dY),Xdzdz=F1o180/(dZ*dZ);
-        const double Edxdx=F1o5040/(dX*dX),Edydy=F1o5040/(dY*dY),Edzdz=F1o5040/(dZ*dZ);
-        const double Sdxdy=F1o4/(dX*dY),Sdxdz=F1o4/(dX*dZ),Sdydz=F1o4/(dY*dZ);
-        const double Fdxdy=F1o144/(dX*dY),Fdxdz=F1o144/(dX*dZ),Fdydz=F1o144/(dY*dZ);
-        const double Xdxdy=F1o3600/(dX*dY),Xdxdz=F1o3600/(dX*dZ),Xdydz=F1o3600/(dY*dZ);
-        const double Edxdy=F1o705600/(dX*dY),Edxdz=F1o705600/(dX*dZ),Edydz=F1o705600/(dY*dZ);
-        const size_t all=(size_t)ex1*ex2*ex3;
-        for(size_t p=0;p<all;++p){fxx[p]=fyy[p]=fzz[p]=fxy[p]=fxz[p]=fyz[p]=ZEO;}
-
-        const int i2_lo=(iminF>0)?iminF:0,j2_lo=(jminF>0)?jminF:0,k2_lo=1,i2_hi=ex1-2,j2_hi=ex2-2,k2_hi=ex3-2;
-        const int i4_lo=(iminF+1>0)?iminF+1:0,j4_lo=(jminF+1>0)?jminF+1:0,k4_lo=2,i4_hi=ex1-3,j4_hi=ex2-3,k4_hi=ex3-3;
-        const int i6_lo=(iminF+2>0)?iminF+2:0,j6_lo=(jminF+2>0)?jminF+2:0,k6_lo=3,i6_hi=ex1-4,j6_hi=ex2-4,k6_hi=ex3-4;
-        const int i8_lo=(iminF+3>0)?iminF+3:0,j8_lo=(jminF+3>0)?jminF+3:0,k8_lo=4,i8_hi=ex1-5,j8_hi=ex2-5,k8_hi=ex3-5;
-        const int has4=(i4_lo<=i4_hi&&j4_lo<=j4_hi&&k4_lo<=k4_hi),has6=(i6_lo<=i6_hi&&j6_lo<=j6_hi&&k6_lo<=k6_hi),has8=(i8_lo<=i8_hi&&j8_lo<=j8_hi&&k8_lo<=k8_hi);
-        #define FH(iF,jF,kF) fh[idx_fh_stbd(iF,jF,kF,ord,ex)]
-
-        /* 2nd-order pass */
-        if(i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi){for(int k0=k2_lo;k0<=k2_hi;++k0){const int kF=k0+1;
-        for(int j0=j2_lo;j0<=j2_hi;++j0){const int jF=j0+1;
-        for(int i0=i2_lo;i0<=i2_hi;++i0){_Bool in4=has4&&i0>=i4_lo&&i0<=i4_hi&&j0>=j4_lo&&j0<=j4_hi&&k0>=k4_lo&&k0<=k4_hi;if(in4)continue;
-            const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fxx[p]=Sdxdx*(FH(iF-1,jF,kF)-TWO*FH(iF,jF,kF)+FH(iF+1,jF,kF));
-            fyy[p]=Sdydy*(FH(iF,jF-1,kF)-TWO*FH(iF,jF,kF)+FH(iF,jF+1,kF));
-            fzz[p]=Sdzdz*(FH(iF,jF,kF-1)-TWO*FH(iF,jF,kF)+FH(iF,jF,kF+1));
-            fxy[p]=Sdxdy*(FH(iF-1,jF-1,kF)-FH(iF+1,jF-1,kF)-FH(iF-1,jF+1,kF)+FH(iF+1,jF+1,kF));
-            fxz[p]=Sdxdz*(FH(iF-1,jF,kF-1)-FH(iF+1,jF,kF-1)-FH(iF-1,jF,kF+1)+FH(iF+1,jF,kF+1));
-            fyz[p]=Sdydz*(FH(iF,jF-1,kF-1)-FH(iF,jF+1,kF-1)-FH(iF,jF-1,kF+1)+FH(iF,jF+1,kF+1));
-        }}}}
-        /* 4th-order pass */
-        if(has4){for(int k0=k4_lo;k0<=k4_hi;++k0){const int kF=k0+1;
-        for(int j0=j4_lo;j0<=j4_hi;++j0){const int jF=j0+1;
-        for(int i0=i4_lo;i0<=i4_hi;++i0){_Bool in6=has6&&i0>=i6_lo&&i0<=i6_hi&&j0>=j6_lo&&j0<=j6_hi&&k0>=k6_lo&&k0<=k6_hi;if(in6)continue;
-            const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fxx[p]=Fdxdx*(-FH(iF-2,jF,kF)+F16*FH(iF-1,jF,kF)-F30*FH(iF,jF,kF)-FH(iF+2,jF,kF)+F16*FH(iF+1,jF,kF));
-            fyy[p]=Fdydy*(-FH(iF,jF-2,kF)+F16*FH(iF,jF-1,kF)-F30*FH(iF,jF,kF)-FH(iF,jF+2,kF)+F16*FH(iF,jF+1,kF));
-            fzz[p]=Fdzdz*(-FH(iF,jF,kF-2)+F16*FH(iF,jF,kF-1)-F30*FH(iF,jF,kF)-FH(iF,jF,kF+2)+F16*FH(iF,jF,kF+1));
-            {const double t_jm2=(FH(iF-2,jF-2,kF)-F8*FH(iF-1,jF-2,kF)+F8*FH(iF+1,jF-2,kF)-FH(iF+2,jF-2,kF));
-             const double t_jm1=(FH(iF-2,jF-1,kF)-F8*FH(iF-1,jF-1,kF)+F8*FH(iF+1,jF-1,kF)-FH(iF+2,jF-1,kF));
-             const double t_jp1=(FH(iF-2,jF+1,kF)-F8*FH(iF-1,jF+1,kF)+F8*FH(iF+1,jF+1,kF)-FH(iF+2,jF+1,kF));
-             const double t_jp2=(FH(iF-2,jF+2,kF)-F8*FH(iF-1,jF+2,kF)+F8*FH(iF+1,jF+2,kF)-FH(iF+2,jF+2,kF));
-             fxy[p]=Fdxdy*(t_jm2-F8*t_jm1+F8*t_jp1-t_jp2);}
-            {const double t_km2=(FH(iF-2,jF,kF-2)-F8*FH(iF-1,jF,kF-2)+F8*FH(iF+1,jF,kF-2)-FH(iF+2,jF,kF-2));
-             const double t_km1=(FH(iF-2,jF,kF-1)-F8*FH(iF-1,jF,kF-1)+F8*FH(iF+1,jF,kF-1)-FH(iF+2,jF,kF-1));
-             const double t_kp1=(FH(iF-2,jF,kF+1)-F8*FH(iF-1,jF,kF+1)+F8*FH(iF+1,jF,kF+1)-FH(iF+2,jF,kF+1));
-             const double t_kp2=(FH(iF-2,jF,kF+2)-F8*FH(iF-1,jF,kF+2)+F8*FH(iF+1,jF,kF+2)-FH(iF+2,jF,kF+2));
-             fxz[p]=Fdxdz*(t_km2-F8*t_km1+F8*t_kp1-t_kp2);}
-            {const double t_km2=(FH(iF,jF-2,kF-2)-F8*FH(iF,jF-1,kF-2)+F8*FH(iF,jF+1,kF-2)-FH(iF,jF+2,kF-2));
-             const double t_km1=(FH(iF,jF-2,kF-1)-F8*FH(iF,jF-1,kF-1)+F8*FH(iF,jF+1,kF-1)-FH(iF,jF+2,kF-1));
-             const double t_kp1=(FH(iF,jF-2,kF+1)-F8*FH(iF,jF-1,kF+1)+F8*FH(iF,jF+1,kF+1)-FH(iF,jF+2,kF+1));
-             const double t_kp2=(FH(iF,jF-2,kF+2)-F8*FH(iF,jF-1,kF+2)+F8*FH(iF,jF+1,kF+2)-FH(iF,jF+2,kF+2));
-             fyz[p]=Fdydz*(t_km2-F8*t_km1+F8*t_kp1-t_kp2);}
-        }}}}
-        /* 6th-order pass */
-        if(has6){for(int k0=k6_lo;k0<=k6_hi;++k0){const int kF=k0+1;
-        for(int j0=j6_lo;j0<=j6_hi;++j0){const int jF=j0+1;
-        for(int i0=i6_lo;i0<=i6_hi;++i0){if(has8&&i0>=i8_lo&&i0<=i8_hi&&j0>=j8_lo&&j0<=j8_hi&&k0>=k8_lo&&k0<=k8_hi)continue;
-            const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fxx[p]=Xdxdx*(TWO*FH(iF-3,jF,kF)-F27*FH(iF-2,jF,kF)+F270*FH(iF-1,jF,kF)-F490*FH(iF,jF,kF)+F270*FH(iF+1,jF,kF)-F27*FH(iF+2,jF,kF)+TWO*FH(iF+3,jF,kF));
-            fyy[p]=Xdydy*(TWO*FH(iF,jF-3,kF)-F27*FH(iF,jF-2,kF)+F270*FH(iF,jF-1,kF)-F490*FH(iF,jF,kF)+F270*FH(iF,jF+1,kF)-F27*FH(iF,jF+2,kF)+TWO*FH(iF,jF+3,kF));
-            fzz[p]=Xdzdz*(TWO*FH(iF,jF,kF-3)-F27*FH(iF,jF,kF-2)+F270*FH(iF,jF,kF-1)-F490*FH(iF,jF,kF)+F270*FH(iF,jF,kF+1)-F27*FH(iF,jF,kF+2)+TWO*FH(iF,jF,kF+3));
-            #define XS6_8(JF,KFDUMMY) (-FH(iF-3,JF,KFDUMMY)+F9*FH(iF-2,JF,KFDUMMY)-F45*FH(iF-1,JF,KFDUMMY)+F45*FH(iF+1,JF,KFDUMMY)-F9*FH(iF+2,JF,KFDUMMY)+FH(iF+3,JF,KFDUMMY))
-            fxy[p]=Xdxdy*(-XS6_8(jF-3,kF)+F9*XS6_8(jF-2,kF)-F45*XS6_8(jF-1,kF)+F45*XS6_8(jF+1,kF)-F9*XS6_8(jF+2,kF)+XS6_8(jF+3,kF));
-            fxz[p]=Xdxdz*(-XS6_8(jF,kF-3)+F9*XS6_8(jF,kF-2)-F45*XS6_8(jF,kF-1)+F45*XS6_8(jF,kF+1)-F9*XS6_8(jF,kF+2)+XS6_8(jF,kF+3));
-            #undef XS6_8
-            #define YS6_8(JF,KFDUMMY) (-FH(iF,JF-3,KFDUMMY)+F9*FH(iF,JF-2,KFDUMMY)-F45*FH(iF,JF-1,KFDUMMY)+F45*FH(iF,JF+1,KFDUMMY)-F9*FH(iF,JF+2,KFDUMMY)+FH(iF,JF+3,KFDUMMY))
-            fyz[p]=Xdydz*(-YS6_8(jF,kF-3)+F9*YS6_8(jF,kF-2)-F45*YS6_8(jF,kF-1)+F45*YS6_8(jF,kF+1)-F9*YS6_8(jF,kF+2)+YS6_8(jF,kF+3));
-            #undef YS6_8
-        }}}}
-        /* 8th-order pass */
-        if(has8){for(int k0=k8_lo;k0<=k8_hi;++k0){const int kF=k0+1;
-        for(int j0=j8_lo;j0<=j8_hi;++j0){const int jF=j0+1;
-        for(int i0=i8_lo;i0<=i8_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fxx[p]=Edxdx*(-(double)9*FH(iF-4,jF,kF)+F128*FH(iF-3,jF,kF)-F1008*FH(iF-2,jF,kF)+F8064*FH(iF-1,jF,kF)-F14350*FH(iF,jF,kF)+F8064*FH(iF+1,jF,kF)-F1008*FH(iF+2,jF,kF)+F128*FH(iF+3,jF,kF)-(double)9*FH(iF+4,jF,kF));
-            fyy[p]=Edydy*(-(double)9*FH(iF,jF-4,kF)+F128*FH(iF,jF-3,kF)-F1008*FH(iF,jF-2,kF)+F8064*FH(iF,jF-1,kF)-F14350*FH(iF,jF,kF)+F8064*FH(iF,jF+1,kF)-F1008*FH(iF,jF+2,kF)+F128*FH(iF,jF+3,kF)-(double)9*FH(iF,jF+4,kF));
-            fzz[p]=Edzdz*(-(double)9*FH(iF,jF,kF-4)+F128*FH(iF,jF,kF-3)-F1008*FH(iF,jF,kF-2)+F8064*FH(iF,jF,kF-1)-F14350*FH(iF,jF,kF)+F8064*FH(iF,jF,kF+1)-F1008*FH(iF,jF,kF+2)+F128*FH(iF,jF,kF+3)-(double)9*FH(iF,jF,kF+4));
-            #define XS8(JF,KFDUMMY) (+(double)3*FH(iF-4,JF,KFDUMMY)-F32*FH(iF-3,JF,KFDUMMY)+F168*FH(iF-2,JF,KFDUMMY)-F672*FH(iF-1,JF,KFDUMMY)+F672*FH(iF+1,JF,KFDUMMY)-F168*FH(iF+2,JF,KFDUMMY)+F32*FH(iF+3,JF,KFDUMMY)-(double)3*FH(iF+4,JF,KFDUMMY))
-            fxy[p]=Edxdy*(+(double)3*XS8(jF-4,kF)-F32*XS8(jF-3,kF)+F168*XS8(jF-2,kF)-F672*XS8(jF-1,kF)+F672*XS8(jF+1,kF)-F168*XS8(jF+2,kF)+F32*XS8(jF+3,kF)-(double)3*XS8(jF+4,kF));
-            fxz[p]=Edxdz*(+(double)3*XS8(jF,kF-4)-F32*XS8(jF,kF-3)+F168*XS8(jF,kF-2)-F672*XS8(jF,kF-1)+F672*XS8(jF,kF+1)-F168*XS8(jF,kF+2)+F32*XS8(jF,kF+3)-(double)3*XS8(jF,kF+4));
-            #undef XS8
-            #define YS8(JF,KFDUMMY) (+(double)3*FH(iF,JF-4,KFDUMMY)-F32*FH(iF,JF-3,KFDUMMY)+F168*FH(iF,JF-2,KFDUMMY)-F672*FH(iF,JF-1,KFDUMMY)+F672*FH(iF,JF+1,KFDUMMY)-F168*FH(iF,JF+2,KFDUMMY)+F32*FH(iF,JF+3,KFDUMMY)-(double)3*FH(iF,JF+4,KFDUMMY))
-            fyz[p]=Edydz*(+(double)3*YS8(jF,kF-4)-F32*YS8(jF,kF-3)+F168*YS8(jF,kF-2)-F672*YS8(jF,kF-1)+F672*YS8(jF,kF+1)-F168*YS8(jF,kF+2)+F32*YS8(jF,kF+3)-(double)3*YS8(jF,kF+4));
-            #undef YS8
-        }}}}
-        #undef FH
-        return;
-    }
-#else
-#error "fdderivs_sh_c.C: unsupported ghost_width"
-#endif
-}

AMSS_NCKU_source/fdderivs_shc_c.C

@@ -1,107 +0,0 @@
-#include "macrodef.h"
-#include "share_func.h"
-#include <cstddef>
-
-/* Forward declarations — Fortran-mangled names from shell C kernels */
-extern "C" {
-void fderivs_sh_(const int ex[3], const double *f,
-                 double *fx, double *fy, double *fz,
-                 const double *X, const double *Y, const double *Z,
-                 double SYM1, double SYM2, double SYM3,
-                 int Symmetry, int onoff, int sst);
-
-void fdderivs_sh_(const int ex[3], const double *f,
-                  double *fxx, double *fxy, double *fxz,
-                  double *fyy, double *fyz, double *fzz,
-                  const double *X, const double *Y, const double *Z,
-                  double SYM1, double SYM2, double SYM3,
-                  int Symmetry, int onoff, int sst);
-
-void fdderivs_shc_(int *ex,
-                    double *f,
-                    double *fxx, double *fxy, double *fxz,
-                    double *fyy, double *fyz, double *fzz,
-                    double *crho, double *sigma, double *R,
-                    double &SYM1, double &SYM2, double &SYM3,
-                    int &Symmetry, int &Lev, int &sst,
-                    double *drhodx, double *drhody, double *drhodz,
-                    double *dsigmadx, double *dsigmady, double *dsigmadz,
-                    double *dRdx, double *dRdy, double *dRdz,
-                    double *drhodxx, double *drhodxy, double *drhodxz,
-                    double *drhodyy, double *drhodyz, double *drhodzz,
-                    double *dsigmadxx, double *dsigmadxy, double *dsigmadxz,
-                    double *dsigmadyy, double *dsigmadyz, double *dsigmadzz,
-                    double *dRdxx, double *dRdxy, double *dRdxz,
-                    double *dRdyy, double *dRdyz, double *dRdzz)
-{
-    const int ex3[3] = { ex[0], ex[1], ex[2] };
-    const size_t n = (size_t)ex[0] * (size_t)ex[1] * (size_t)ex[2];
-
-    double *gx  = (double*)malloc(n * sizeof(double));
-    double *gy  = (double*)malloc(n * sizeof(double));
-    double *gz  = (double*)malloc(n * sizeof(double));
-    double *gxx = (double*)malloc(n * sizeof(double));
-    double *gxy = (double*)malloc(n * sizeof(double));
-    double *gxz = (double*)malloc(n * sizeof(double));
-    double *gyy = (double*)malloc(n * sizeof(double));
-    double *gyz = (double*)malloc(n * sizeof(double));
-    double *gzz = (double*)malloc(n * sizeof(double));
-
-    if (!gx||!gy||!gz||!gxx||!gxy||!gxz||!gyy||!gyz||!gzz) {
-        free(gx);free(gy);free(gz);free(gxx);free(gxy);free(gxz);free(gyy);free(gyz);free(gzz);
-        return;
-    }
-
-    fderivs_sh_(ex3, f, gx, gy, gz, crho, sigma, R, SYM1, SYM2, SYM3, Symmetry, Lev, sst);
-    fdderivs_sh_(ex3, f, gxx, gxy, gxz, gyy, gyz, gzz, crho, sigma, R, SYM1, SYM2, SYM3, Symmetry, Lev, sst);
-
-    for (size_t i = 0; i < n; ++i) {
-        const double rx=drhodx[i], ry=drhody[i], rz=drhodz[i];
-        const double sx=dsigmadx[i], sy=dsigmady[i], sz=dsigmadz[i];
-        const double Rx=dRdx[i], Ry=dRdy[i], Rz=dRdz[i];
-        const double rxx=drhodxx[i], rxy=drhodxy[i], rxz=drhodxz[i];
-        const double ryy=drhodyy[i], ryz=drhodyz[i], rzz=drhodzz[i];
-        const double sxx=dsigmadxx[i], sxy=dsigmadxy[i], sxz=dsigmadxz[i];
-        const double syy=dsigmadyy[i], syz=dsigmadyz[i], szz=dsigmadzz[i];
-        const double Rxx=dRdxx[i], Rxy=dRdxy[i], Rxz=dRdxz[i];
-        const double Ryy=dRdyy[i], Ryz=dRdyz[i], Rzz=dRdzz[i];
-
-        const double Gr=gx[i], Gs=gy[i], GR=gz[i];
-        const double Grr=gxx[i], Grs=gxy[i], GrR=gxz[i];
-        const double Gss=gyy[i], GsR=gyz[i], GRR=gzz[i];
-
-        /* fxx */
-        fxx[i] = rx*rx*Grr + sx*sx*Gss + Rx*Rx*GRR
-               + 2.0*(rx*sx*Grs + rx*Rx*GrR + sx*Rx*GsR)
-               + rxx*Gr + sxx*Gs + Rxx*GR;
-
-        /* fxy */
-        fxy[i] = rx*ry*Grr + sx*sy*Gss + Rx*Ry*GRR
-               + rx*sy*Grs + ry*sx*Grs + rx*Ry*GrR + ry*Rx*GrR + sx*Ry*GsR + sy*Rx*GsR
-               + rxy*Gr + sxy*Gs + Rxy*GR;
-
-        /* fxz */
-        fxz[i] = rx*rz*Grr + sx*sz*Gss + Rx*Rz*GRR
-               + rx*sz*Grs + rz*sx*Grs + rx*Rz*GrR + rz*Rx*GrR + sx*Rz*GsR + sz*Rx*GsR
-               + rxz*Gr + sxz*Gs + Rxz*GR;
-
-        /* fyy */
-        fyy[i] = ry*ry*Grr + sy*sy*Gss + Ry*Ry*GRR
-               + 2.0*(ry*sy*Grs + ry*Ry*GrR + sy*Ry*GsR)
-               + ryy*Gr + syy*Gs + Ryy*GR;
-
-        /* fyz */
-        fyz[i] = ry*rz*Grr + sy*sz*Gss + Ry*Rz*GRR
-               + ry*sz*Grs + rz*sy*Grs + ry*Rz*GrR + rz*Ry*GrR + sy*Rz*GsR + sz*Ry*GsR
-               + ryz*Gr + syz*Gs + Ryz*GR;
-
-        /* fzz */
-        fzz[i] = rz*rz*Grr + sz*sz*Gss + Rz*Rz*GRR
-               + 2.0*(rz*sz*Grs + rz*Rz*GrR + sz*Rz*GsR)
-               + rzz*Gr + szz*Gs + Rzz*GR;
-    }
-
-    free(gx);free(gy);free(gz);free(gxx);free(gxy);free(gxz);free(gyy);free(gyz);free(gzz);
-}
-
-} // extern "C"

AMSS_NCKU_source/fderivs_c.C

@@ -1,18 +1,14 @@
-#include "macrodef.h"
 #include "tool.h"
 
 /*
- * C 版 fderivs — first derivatives df/dx, df/dy, df/dz.
+ * C 版 fderivs
  *
- * Finite difference order is selected at compile time via the ghost_width macro
- * (defined in macrodef.fh):
- *   ghost_width = 2 → 2nd-order
- *   ghost_width = 3 → 4th-order
- *   ghost_width = 4 → 6th-order
- *   ghost_width = 5 → 8th-order
+ * Fortran:
+ * subroutine fderivs(ex,f,fx,fy,fz,X,Y,Z,SYM1,SYM2,SYM3,symmetry,onoff)
  *
- * Multi-pass overwrite strategy: compute the widest (lowest-order) stencil first,
- * then overwrite interior regions with progressively higher-order stencils.
+ * 约定：
+ *   f, fx, fy, fz: ex1*ex2*ex3，按 idx_ex 布局
+ *   X: ex1, Y: ex2, Z: ex3
  */
 void fderivs(const int ex[3],
              const double *f,
@@ -21,596 +17,151 @@ void fderivs(const int ex[3],
              double SYM1, double SYM2, double SYM3,
              int Symmetry, int onoff)
 {
-    (void)onoff;
+    (void)onoff; // Fortran 里没用到
 
-    const double ZEO = 0.0, ONE = 1.0, TWO = 2.0, EIT = 8.0;
-    const double F9 = 9.0, F12 = 12.0, F45 = 45.0, F60 = 60.0;
-    const double F32 = 32.0, F168 = 168.0, F672 = 672.0, F840 = 840.0;
+    const double ZEO = 0.0, ONE = 1.0;
+    const double TWO = 2.0, EIT = 8.0;
+    const double F12 = 12.0;
 
-    const int NO_SYMM = 0, EQ_SYMM = 1;
+    const int NO_SYMM = 0, EQ_SYMM = 1; // OCTANT=2 在本子程序里不直接用
 
     const int ex1 = ex[0], ex2 = ex[1], ex3 = ex[2];
+
+    // dX = X(2)-X(1) -> C: X[1]-X[0]
     const double dX = X[1] - X[0];
     const double dY = Y[1] - Y[0];
     const double dZ = Z[1] - Z[0];
 
-    const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
+    // Fortran 1-based bounds
+    const int imaxF = ex1;
+    const int jmaxF = ex2;
+    const int kmaxF = ex3;
 
-    const int gw = ghost_width; // compile-time constant
+    int iminF = 1, jminF = 1, kminF = 1;
+    if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -1;
+    if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -1;
+    if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -1;
 
-#if (ghost_width == 2)
-    /* ---- 2nd-order ------------------------------------------------------ */
-    {
-        const int ord = 1; // symmetry_bd ord = ghost_width - 1
+    // SoA(1:3) = SYM1,SYM2,SYM3
+    const double SoA[3] = { SYM1, SYM2, SYM3 };
 
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = 0;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = 0;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = 0;
+    // fh: (ex1+2)*(ex2+2)*(ex3+2) because ord=2
+    const size_t nx = (size_t)ex1 + 2;
+    const size_t ny = (size_t)ex2 + 2;
+    const size_t nz = (size_t)ex3 + 2;
+    const size_t fh_size = nx * ny * nz;
+    static double *fh = NULL;
+    static size_t cap = 0;
 
-        const double SoA[3] = { SYM1, SYM2, SYM3 };
-
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
-
-        static double *fh_buf = NULL;
-        static size_t cap = 0;
-        if (fh_size > cap) {
-            free(fh_buf);
-            fh_buf = (double*)aligned_alloc(64, fh_size * sizeof(double));
-            cap = fh_size;
-        }
-        double *fh = fh_buf;
-        if (!fh) return;
-
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d2dx = ONE / TWO / dX;
-        const double d2dy = ONE / TWO / dY;
-        const double d2dz = ONE / TWO / dZ;
-
-        const size_t all = (size_t)ex1 * (size_t)ex2 * (size_t)ex3;
-        for (size_t p = 0; p < all; ++p) {
-            fx[p] = ZEO; fy[p] = ZEO; fz[p] = ZEO;
-        }
-
-        /* 2nd-order pass: [-1, 0, +1] / (2*dx) */
-        const int i2_lo = (iminF > 0) ? iminF : 0;
-        const int j2_lo = (jminF > 0) ? jminF : 0;
-        const int k2_lo = (kminF > 0) ? kminF : 0;
-        const int i2_hi = ex1 - 2;
-        const int j2_hi = ex2 - 2;
-        const int k2_hi = ex3 - 2;
-
-        if (i2_lo <= i2_hi && j2_lo <= j2_hi && k2_lo <= k2_hi) {
-            for (int k0 = k2_lo; k0 <= k2_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j2_lo; j0 <= j2_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i2_lo; i0 <= i2_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d2dx * (
-                            -fh[idx_fh_F_ord1(iF - 1, jF,     kF,     ex)] +
-                             fh[idx_fh_F_ord1(iF + 1, jF,     kF,     ex)]
-                        );
-
-                        fy[p] = d2dy * (
-                            -fh[idx_fh_F_ord1(iF,     jF - 1, kF,     ex)] +
-                             fh[idx_fh_F_ord1(iF,     jF + 1, kF,     ex)]
-                        );
-
-                        fz[p] = d2dz * (
-                            -fh[idx_fh_F_ord1(iF,     jF,     kF - 1, ex)] +
-                             fh[idx_fh_F_ord1(iF,     jF,     kF + 1, ex)]
-                        );
-                    }
-                }
-            }
-        }
-        return;
+    if (fh_size > cap) {
+        free(fh);
+        fh = (double*)aligned_alloc(64, fh_size * sizeof(double));
+        cap = fh_size;
     }
-#elif (ghost_width == 3)
-    /* ---- 4th-order (original code) ------------------------------------ */
-    {
-        const int ord = 2; // symmetry_bd ord
+    // double *fh = (double*)malloc(fh_size * sizeof(double));
+    if (!fh) return;
 
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -1;
+    // call symmetry_bd(2,ex,f,fh,SoA)
+    symmetry_bd(2, ex, f, fh, SoA);
 
-        const double SoA[3] = { SYM1, SYM2, SYM3 };
+    const double d12dx = ONE / F12 / dX;
+    const double d12dy = ONE / F12 / dY;
+    const double d12dz = ONE / F12 / dZ;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
+    const double d2dx  = ONE / TWO / dX;
+    const double d2dy  = ONE / TWO / dY;
+    const double d2dz  = ONE / TWO / dZ;
 
-        static double *fh_buf = NULL;
-        static size_t cap = 0;
-        if (fh_size > cap) {
-            free(fh_buf);
-            fh_buf = (double*)aligned_alloc(64, fh_size * sizeof(double));
-            cap = fh_size;
-        }
-        double *fh = fh_buf;
-        if (!fh) return;
-
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d12dx = ONE / F12 / dX;
-        const double d12dy = ONE / F12 / dY;
-        const double d12dz = ONE / F12 / dZ;
-        const double d2dx  = ONE / TWO / dX;
-        const double d2dy  = ONE / TWO / dY;
-        const double d2dz  = ONE / TWO / dZ;
-
-        const size_t all = (size_t)ex1 * (size_t)ex2 * (size_t)ex3;
-        for (size_t p = 0; p < all; ++p) {
-            fx[p] = ZEO; fy[p] = ZEO; fz[p] = ZEO;
-        }
-
-        const int i2_lo = (iminF > 0) ? iminF : 0;
-        const int j2_lo = (jminF > 0) ? jminF : 0;
-        const int k2_lo = (kminF > 0) ? kminF : 0;
-        const int i2_hi = ex1 - 2;
-        const int j2_hi = ex2 - 2;
-        const int k2_hi = ex3 - 2;
-
-        const int i4_lo = (iminF + 1 > 0) ? (iminF + 1) : 0;
-        const int j4_lo = (jminF + 1 > 0) ? (jminF + 1) : 0;
-        const int k4_lo = (kminF + 1 > 0) ? (kminF + 1) : 0;
-        const int i4_hi = ex1 - 3;
-        const int j4_hi = ex2 - 3;
-        const int k4_hi = ex3 - 3;
-
-        if (i2_lo <= i2_hi && j2_lo <= j2_hi && k2_lo <= k2_hi) {
-            for (int k0 = k2_lo; k0 <= k2_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j2_lo; j0 <= j2_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i2_lo; i0 <= i2_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d2dx * (
-                            -fh[idx_fh_F_ord2(iF - 1, jF,     kF,     ex)] +
-                             fh[idx_fh_F_ord2(iF + 1, jF,     kF,     ex)]
-                        );
-
-                        fy[p] = d2dy * (
-                            -fh[idx_fh_F_ord2(iF,     jF - 1, kF,     ex)] +
-                             fh[idx_fh_F_ord2(iF,     jF + 1, kF,     ex)]
-                        );
-
-                        fz[p] = d2dz * (
-                            -fh[idx_fh_F_ord2(iF,     jF,     kF - 1, ex)] +
-                             fh[idx_fh_F_ord2(iF,     jF,     kF + 1, ex)]
-                        );
-                    }
-                }
-            }
-        }
-
-        if (i4_lo <= i4_hi && j4_lo <= j4_hi && k4_lo <= k4_hi) {
-            for (int k0 = k4_lo; k0 <= k4_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j4_lo; j0 <= j4_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i4_lo; i0 <= i4_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d12dx * (
-                            fh[idx_fh_F_ord2(iF - 2, jF,     kF,     ex)] -
-                            EIT * fh[idx_fh_F_ord2(iF - 1, jF,     kF,     ex)] +
-                            EIT * fh[idx_fh_F_ord2(iF + 1, jF,     kF,     ex)] -
-                            fh[idx_fh_F_ord2(iF + 2, jF,     kF,     ex)]
-                        );
-
-                        fy[p] = d12dy * (
-                            fh[idx_fh_F_ord2(iF,     jF - 2, kF,     ex)] -
-                            EIT * fh[idx_fh_F_ord2(iF,     jF - 1, kF,     ex)] +
-                            EIT * fh[idx_fh_F_ord2(iF,     jF + 1, kF,     ex)] -
-                            fh[idx_fh_F_ord2(iF,     jF + 2, kF,     ex)]
-                        );
-
-                        fz[p] = d12dz * (
-                            fh[idx_fh_F_ord2(iF,     jF,     kF - 2, ex)] -
-                            EIT * fh[idx_fh_F_ord2(iF,     jF,     kF - 1, ex)] +
-                            EIT * fh[idx_fh_F_ord2(iF,     jF,     kF + 1, ex)] -
-                            fh[idx_fh_F_ord2(iF,     jF,     kF + 2, ex)]
-                        );
-                    }
-                }
-            }
-        }
-        return;
+    // fx = fy = fz = 0
+    const size_t all = (size_t)ex1 * (size_t)ex2 * (size_t)ex3;
+    for (size_t p = 0; p < all; ++p) {
+        fx[p] = ZEO;
+        fy[p] = ZEO;
+        fz[p] = ZEO;
     }
-#elif (ghost_width == 4)
-    /* ---- 6th-order ----------------------------------------------------- */
-    {
-        const int ord = 3;
 
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -2;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -2;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -2;
+    /*
+     * 两段式：
+     * 1) 先在二阶可用区域计算二阶模板
+     * 2) 再在高阶可用区域覆盖为四阶模板
+     *
+     * 与原 if/elseif 逻辑等价，但减少逐点分支判断。
+     */
+    const int i2_lo = (iminF > 0) ? iminF : 0;
+    const int j2_lo = (jminF > 0) ? jminF : 0;
+    const int k2_lo = (kminF > 0) ? kminF : 0;
+    const int i2_hi = ex1 - 2;
+    const int j2_hi = ex2 - 2;
+    const int k2_hi = ex3 - 2;
 
-        const double SoA[3] = { SYM1, SYM2, SYM3 };
+    const int i4_lo = (iminF + 1 > 0) ? (iminF + 1) : 0;
+    const int j4_lo = (jminF + 1 > 0) ? (jminF + 1) : 0;
+    const int k4_lo = (kminF + 1 > 0) ? (kminF + 1) : 0;
+    const int i4_hi = ex1 - 3;
+    const int j4_hi = ex2 - 3;
+    const int k4_hi = ex3 - 3;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
+    if (i2_lo <= i2_hi && j2_lo <= j2_hi && k2_lo <= k2_hi) {
+        for (int k0 = k2_lo; k0 <= k2_hi; ++k0) {
+            const int kF = k0 + 1;
+            for (int j0 = j2_lo; j0 <= j2_hi; ++j0) {
+                const int jF = j0 + 1;
+                for (int i0 = i2_lo; i0 <= i2_hi; ++i0) {
+                    const int iF = i0 + 1;
+                    const size_t p = idx_ex(i0, j0, k0, ex);
 
-        static double *fh_buf = NULL;
-        static size_t cap = 0;
-        if (fh_size > cap) {
-            free(fh_buf);
-            fh_buf = (double*)aligned_alloc(64, fh_size * sizeof(double));
-            cap = fh_size;
-        }
-        double *fh = fh_buf;
-        if (!fh) return;
+                    fx[p] = d2dx * (
+                        -fh[idx_fh_F_ord2(iF - 1, jF,     kF,     ex)] +
+                         fh[idx_fh_F_ord2(iF + 1, jF,     kF,     ex)]
+                    );
 
-        symmetry_bd(ord, ex, f, fh, SoA);
+                    fy[p] = d2dy * (
+                        -fh[idx_fh_F_ord2(iF,     jF - 1, kF,     ex)] +
+                         fh[idx_fh_F_ord2(iF,     jF + 1, kF,     ex)]
+                    );
 
-        /* Denominators */
-        const double d60dx = ONE / F60 / dX;
-        const double d60dy = ONE / F60 / dY;
-        const double d60dz = ONE / F60 / dZ;
-        const double d12dx = ONE / F12 / dX;
-        const double d12dy = ONE / F12 / dY;
-        const double d12dz = ONE / F12 / dZ;
-        const double d2dx  = ONE / TWO / dX;
-        const double d2dy  = ONE / TWO / dY;
-        const double d2dz  = ONE / TWO / dZ;
-
-        const size_t all = (size_t)ex1 * (size_t)ex2 * (size_t)ex3;
-        for (size_t p = 0; p < all; ++p) {
-            fx[p] = ZEO; fy[p] = ZEO; fz[p] = ZEO;
-        }
-
-        /* 2nd-order pass: 3pt, widest */
-        const int i2_lo = (iminF > 0) ? iminF : 0;
-        const int j2_lo = (jminF > 0) ? jminF : 0;
-        const int k2_lo = (kminF > 0) ? kminF : 0;
-        const int i2_hi = ex1 - 2;
-        const int j2_hi = ex2 - 2;
-        const int k2_hi = ex3 - 2;
-
-        /* 4th-order pass: 5pt */
-        const int i4_lo = (iminF + 1 > 0) ? (iminF + 1) : 0;
-        const int j4_lo = (jminF + 1 > 0) ? (jminF + 1) : 0;
-        const int k4_lo = (kminF + 1 > 0) ? (kminF + 1) : 0;
-        const int i4_hi = ex1 - 3;
-        const int j4_hi = ex2 - 3;
-        const int k4_hi = ex3 - 3;
-
-        /* 6th-order pass: 7pt, narrowest interior */
-        const int i6_lo = (iminF + 2 > 0) ? (iminF + 2) : 0;
-        const int j6_lo = (jminF + 2 > 0) ? (jminF + 2) : 0;
-        const int k6_lo = (kminF + 2 > 0) ? (kminF + 2) : 0;
-        const int i6_hi = ex1 - 4;
-        const int j6_hi = ex2 - 4;
-        const int k6_hi = ex3 - 4;
-
-        /* 2nd-order */
-        if (i2_lo <= i2_hi && j2_lo <= j2_hi && k2_lo <= k2_hi) {
-            for (int k0 = k2_lo; k0 <= k2_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j2_lo; j0 <= j2_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i2_lo; i0 <= i2_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d2dx * (
-                            -fh[idx_fh_F(iF - 1, jF,     kF,     ex)] +
-                             fh[idx_fh_F(iF + 1, jF,     kF,     ex)]);
-                        fy[p] = d2dy * (
-                            -fh[idx_fh_F(iF,     jF - 1, kF,     ex)] +
-                             fh[idx_fh_F(iF,     jF + 1, kF,     ex)]);
-                        fz[p] = d2dz * (
-                            -fh[idx_fh_F(iF,     jF,     kF - 1, ex)] +
-                             fh[idx_fh_F(iF,     jF,     kF + 1, ex)]);
-                    }
+                    fz[p] = d2dz * (
+                        -fh[idx_fh_F_ord2(iF,     jF,     kF - 1, ex)] +
+                         fh[idx_fh_F_ord2(iF,     jF,     kF + 1, ex)]
+                    );
                 }
             }
         }
-
-        /* 4th-order overwrite */
-        if (i4_lo <= i4_hi && j4_lo <= j4_hi && k4_lo <= k4_hi) {
-            for (int k0 = k4_lo; k0 <= k4_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j4_lo; j0 <= j4_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i4_lo; i0 <= i4_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d12dx * (
-                            fh[idx_fh_F(iF - 2, jF,     kF,     ex)] -
-                            EIT * fh[idx_fh_F(iF - 1, jF,     kF,     ex)] +
-                            EIT * fh[idx_fh_F(iF + 1, jF,     kF,     ex)] -
-                            fh[idx_fh_F(iF + 2, jF,     kF,     ex)]);
-
-                        fy[p] = d12dy * (
-                            fh[idx_fh_F(iF,     jF - 2, kF,     ex)] -
-                            EIT * fh[idx_fh_F(iF,     jF - 1, kF,     ex)] +
-                            EIT * fh[idx_fh_F(iF,     jF + 1, kF,     ex)] -
-                            fh[idx_fh_F(iF,     jF + 2, kF,     ex)]);
-
-                        fz[p] = d12dz * (
-                            fh[idx_fh_F(iF,     jF,     kF - 2, ex)] -
-                            EIT * fh[idx_fh_F(iF,     jF,     kF - 1, ex)] +
-                            EIT * fh[idx_fh_F(iF,     jF,     kF + 1, ex)] -
-                            fh[idx_fh_F(iF,     jF,     kF + 2, ex)]);
-                    }
-                }
-            }
-        }
-
-        /* 6th-order overwrite: [-1,+9,-45,0,+45,-9,+1] / (60*dx) */
-        if (i6_lo <= i6_hi && j6_lo <= j6_hi && k6_lo <= k6_hi) {
-            for (int k0 = k6_lo; k0 <= k6_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j6_lo; j0 <= j6_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i6_lo; i0 <= i6_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d60dx * (
-                            -fh[idx_fh_F(iF - 3, jF,     kF,     ex)] +
-                            F9 * fh[idx_fh_F(iF - 2, jF,     kF,     ex)] -
-                            F45 * fh[idx_fh_F(iF - 1, jF,     kF,     ex)] +
-                            F45 * fh[idx_fh_F(iF + 1, jF,     kF,     ex)] -
-                            F9 * fh[idx_fh_F(iF + 2, jF,     kF,     ex)] +
-                            fh[idx_fh_F(iF + 3, jF,     kF,     ex)]);
-
-                        fy[p] = d60dy * (
-                            -fh[idx_fh_F(iF,     jF - 3, kF,     ex)] +
-                            F9 * fh[idx_fh_F(iF,     jF - 2, kF,     ex)] -
-                            F45 * fh[idx_fh_F(iF,     jF - 1, kF,     ex)] +
-                            F45 * fh[idx_fh_F(iF,     jF + 1, kF,     ex)] -
-                            F9 * fh[idx_fh_F(iF,     jF + 2, kF,     ex)] +
-                            fh[idx_fh_F(iF,     jF + 3, kF,     ex)]);
-
-                        fz[p] = d60dz * (
-                            -fh[idx_fh_F(iF,     jF,     kF - 3, ex)] +
-                            F9 * fh[idx_fh_F(iF,     jF,     kF - 2, ex)] -
-                            F45 * fh[idx_fh_F(iF,     jF,     kF - 1, ex)] +
-                            F45 * fh[idx_fh_F(iF,     jF,     kF + 1, ex)] -
-                            F9 * fh[idx_fh_F(iF,     jF,     kF + 2, ex)] +
-                            fh[idx_fh_F(iF,     jF,     kF + 3, ex)]);
-                    }
-                }
-            }
-        }
-        return;
     }
-#elif (ghost_width == 5)
-    /* ---- 8th-order ----------------------------------------------------- */
-    {
-        const int ord = 4;
 
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -3;
+    if (i4_lo <= i4_hi && j4_lo <= j4_hi && k4_lo <= k4_hi) {
+        for (int k0 = k4_lo; k0 <= k4_hi; ++k0) {
+            const int kF = k0 + 1;
+            for (int j0 = j4_lo; j0 <= j4_hi; ++j0) {
+                const int jF = j0 + 1;
+                for (int i0 = i4_lo; i0 <= i4_hi; ++i0) {
+                    const int iF = i0 + 1;
+                    const size_t p = idx_ex(i0, j0, k0, ex);
 
-        const double SoA[3] = { SYM1, SYM2, SYM3 };
+                    fx[p] = d12dx * (
+                        fh[idx_fh_F_ord2(iF - 2, jF,     kF,     ex)] -
+                        EIT * fh[idx_fh_F_ord2(iF - 1, jF,     kF,     ex)] +
+                        EIT * fh[idx_fh_F_ord2(iF + 1, jF,     kF,     ex)] -
+                        fh[idx_fh_F_ord2(iF + 2, jF,     kF,     ex)]
+                    );
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
+                    fy[p] = d12dy * (
+                        fh[idx_fh_F_ord2(iF,     jF - 2, kF,     ex)] -
+                        EIT * fh[idx_fh_F_ord2(iF,     jF - 1, kF,     ex)] +
+                        EIT * fh[idx_fh_F_ord2(iF,     jF + 1, kF,     ex)] -
+                        fh[idx_fh_F_ord2(iF,     jF + 2, kF,     ex)]
+                    );
 
-        static double *fh_buf = NULL;
-        static size_t cap = 0;
-        if (fh_size > cap) {
-            free(fh_buf);
-            fh_buf = (double*)aligned_alloc(64, fh_size * sizeof(double));
-            cap = fh_size;
-        }
-        double *fh = fh_buf;
-        if (!fh) return;
-
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d840dx = ONE / F840 / dX;
-        const double d840dy = ONE / F840 / dY;
-        const double d840dz = ONE / F840 / dZ;
-        const double d60dx  = ONE / F60  / dX;
-        const double d60dy  = ONE / F60  / dY;
-        const double d60dz  = ONE / F60  / dZ;
-        const double d12dx  = ONE / F12  / dX;
-        const double d12dy  = ONE / F12  / dY;
-        const double d12dz  = ONE / F12  / dZ;
-        const double d2dx   = ONE / TWO  / dX;
-        const double d2dy   = ONE / TWO  / dY;
-        const double d2dz   = ONE / TWO  / dZ;
-
-        const size_t all = (size_t)ex1 * (size_t)ex2 * (size_t)ex3;
-        for (size_t p = 0; p < all; ++p) {
-            fx[p] = ZEO; fy[p] = ZEO; fz[p] = ZEO;
-        }
-
-        /* 2nd: 3pt, widest */
-        const int i2_lo = (iminF > 0) ? iminF : 0;
-        const int j2_lo = (jminF > 0) ? jminF : 0;
-        const int k2_lo = (kminF > 0) ? kminF : 0;
-        const int i2_hi = ex1 - 2;
-        const int j2_hi = ex2 - 2;
-        const int k2_hi = ex3 - 2;
-
-        /* 4th: 5pt */
-        const int i4_lo = (iminF + 1 > 0) ? (iminF + 1) : 0;
-        const int j4_lo = (jminF + 1 > 0) ? (jminF + 1) : 0;
-        const int k4_lo = (kminF + 1 > 0) ? (kminF + 1) : 0;
-        const int i4_hi = ex1 - 3;
-        const int j4_hi = ex2 - 3;
-        const int k4_hi = ex3 - 3;
-
-        /* 6th: 7pt */
-        const int i6_lo = (iminF + 2 > 0) ? (iminF + 2) : 0;
-        const int j6_lo = (jminF + 2 > 0) ? (jminF + 2) : 0;
-        const int k6_lo = (kminF + 2 > 0) ? (kminF + 2) : 0;
-        const int i6_hi = ex1 - 4;
-        const int j6_hi = ex2 - 4;
-        const int k6_hi = ex3 - 4;
-
-        /* 8th: 9pt, narrowest */
-        const int i8_lo = (iminF + 3 > 0) ? (iminF + 3) : 0;
-        const int j8_lo = (jminF + 3 > 0) ? (jminF + 3) : 0;
-        const int k8_lo = (kminF + 3 > 0) ? (kminF + 3) : 0;
-        const int i8_hi = ex1 - 5;
-        const int j8_hi = ex2 - 5;
-        const int k8_hi = ex3 - 5;
-
-        if (i2_lo <= i2_hi && j2_lo <= j2_hi && k2_lo <= k2_hi) {
-            for (int k0 = k2_lo; k0 <= k2_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j2_lo; j0 <= j2_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i2_lo; i0 <= i2_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d2dx * (
-                            -fh[idx_fh_F_ord4(iF - 1, jF,     kF,     ex)] +
-                             fh[idx_fh_F_ord4(iF + 1, jF,     kF,     ex)]);
-                        fy[p] = d2dy * (
-                            -fh[idx_fh_F_ord4(iF,     jF - 1, kF,     ex)] +
-                             fh[idx_fh_F_ord4(iF,     jF + 1, kF,     ex)]);
-                        fz[p] = d2dz * (
-                            -fh[idx_fh_F_ord4(iF,     jF,     kF - 1, ex)] +
-                             fh[idx_fh_F_ord4(iF,     jF,     kF + 1, ex)]);
-                    }
+                    fz[p] = d12dz * (
+                        fh[idx_fh_F_ord2(iF,     jF,     kF - 2, ex)] -
+                        EIT * fh[idx_fh_F_ord2(iF,     jF,     kF - 1, ex)] +
+                        EIT * fh[idx_fh_F_ord2(iF,     jF,     kF + 1, ex)] -
+                        fh[idx_fh_F_ord2(iF,     jF,     kF + 2, ex)]
+                    );
                 }
             }
         }
-
-        if (i4_lo <= i4_hi && j4_lo <= j4_hi && k4_lo <= k4_hi) {
-            for (int k0 = k4_lo; k0 <= k4_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j4_lo; j0 <= j4_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i4_lo; i0 <= i4_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d12dx * (
-                            fh[idx_fh_F_ord4(iF - 2, jF,     kF,     ex)] -
-                            EIT * fh[idx_fh_F_ord4(iF - 1, jF,     kF,     ex)] +
-                            EIT * fh[idx_fh_F_ord4(iF + 1, jF,     kF,     ex)] -
-                            fh[idx_fh_F_ord4(iF + 2, jF,     kF,     ex)]);
-
-                        fy[p] = d12dy * (
-                            fh[idx_fh_F_ord4(iF,     jF - 2, kF,     ex)] -
-                            EIT * fh[idx_fh_F_ord4(iF,     jF - 1, kF,     ex)] +
-                            EIT * fh[idx_fh_F_ord4(iF,     jF + 1, kF,     ex)] -
-                            fh[idx_fh_F_ord4(iF,     jF + 2, kF,     ex)]);
-
-                        fz[p] = d12dz * (
-                            fh[idx_fh_F_ord4(iF,     jF,     kF - 2, ex)] -
-                            EIT * fh[idx_fh_F_ord4(iF,     jF,     kF - 1, ex)] +
-                            EIT * fh[idx_fh_F_ord4(iF,     jF,     kF + 1, ex)] -
-                            fh[idx_fh_F_ord4(iF,     jF,     kF + 2, ex)]);
-                    }
-                }
-            }
-        }
-
-        if (i6_lo <= i6_hi && j6_lo <= j6_hi && k6_lo <= k6_hi) {
-            for (int k0 = k6_lo; k0 <= k6_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j6_lo; j0 <= j6_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i6_lo; i0 <= i6_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d60dx * (
-                            -fh[idx_fh_F_ord4(iF - 3, jF,     kF,     ex)] +
-                            F9 * fh[idx_fh_F_ord4(iF - 2, jF,     kF,     ex)] -
-                            F45 * fh[idx_fh_F_ord4(iF - 1, jF,     kF,     ex)] +
-                            F45 * fh[idx_fh_F_ord4(iF + 1, jF,     kF,     ex)] -
-                            F9 * fh[idx_fh_F_ord4(iF + 2, jF,     kF,     ex)] +
-                            fh[idx_fh_F_ord4(iF + 3, jF,     kF,     ex)]);
-
-                        fy[p] = d60dy * (
-                            -fh[idx_fh_F_ord4(iF,     jF - 3, kF,     ex)] +
-                            F9 * fh[idx_fh_F_ord4(iF,     jF - 2, kF,     ex)] -
-                            F45 * fh[idx_fh_F_ord4(iF,     jF - 1, kF,     ex)] +
-                            F45 * fh[idx_fh_F_ord4(iF,     jF + 1, kF,     ex)] -
-                            F9 * fh[idx_fh_F_ord4(iF,     jF + 2, kF,     ex)] +
-                            fh[idx_fh_F_ord4(iF,     jF + 3, kF,     ex)]);
-
-                        fz[p] = d60dz * (
-                            -fh[idx_fh_F_ord4(iF,     jF,     kF - 3, ex)] +
-                            F9 * fh[idx_fh_F_ord4(iF,     jF,     kF - 2, ex)] -
-                            F45 * fh[idx_fh_F_ord4(iF,     jF,     kF - 1, ex)] +
-                            F45 * fh[idx_fh_F_ord4(iF,     jF,     kF + 1, ex)] -
-                            F9 * fh[idx_fh_F_ord4(iF,     jF,     kF + 2, ex)] +
-                            fh[idx_fh_F_ord4(iF,     jF,     kF + 3, ex)]);
-                    }
-                }
-            }
-        }
-
-        /* 8th-order overwrite: [+3,-32,+168,-672,0,+672,-168,+32,-3] / (840*dx) */
-        if (i8_lo <= i8_hi && j8_lo <= j8_hi && k8_lo <= k8_hi) {
-            for (int k0 = k8_lo; k0 <= k8_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j8_lo; j0 <= j8_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i8_lo; i0 <= i8_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        fx[p] = d840dx * (
-                            +(double)3 * fh[idx_fh_F_ord4(iF - 4, jF,     kF,     ex)] -
-                            F32  * fh[idx_fh_F_ord4(iF - 3, jF,     kF,     ex)] +
-                            F168 * fh[idx_fh_F_ord4(iF - 2, jF,     kF,     ex)] -
-                            F672 * fh[idx_fh_F_ord4(iF - 1, jF,     kF,     ex)] +
-                            F672 * fh[idx_fh_F_ord4(iF + 1, jF,     kF,     ex)] -
-                            F168 * fh[idx_fh_F_ord4(iF + 2, jF,     kF,     ex)] +
-                            F32  * fh[idx_fh_F_ord4(iF + 3, jF,     kF,     ex)] -
-                            (double)3 * fh[idx_fh_F_ord4(iF + 4, jF,     kF,     ex)]);
-
-                        fy[p] = d840dy * (
-                            +(double)3 * fh[idx_fh_F_ord4(iF,     jF - 4, kF,     ex)] -
-                            F32  * fh[idx_fh_F_ord4(iF,     jF - 3, kF,     ex)] +
-                            F168 * fh[idx_fh_F_ord4(iF,     jF - 2, kF,     ex)] -
-                            F672 * fh[idx_fh_F_ord4(iF,     jF - 1, kF,     ex)] +
-                            F672 * fh[idx_fh_F_ord4(iF,     jF + 1, kF,     ex)] -
-                            F168 * fh[idx_fh_F_ord4(iF,     jF + 2, kF,     ex)] +
-                            F32  * fh[idx_fh_F_ord4(iF,     jF + 3, kF,     ex)] -
-                            (double)3 * fh[idx_fh_F_ord4(iF,     jF + 4, kF,     ex)]);
-
-                        fz[p] = d840dz * (
-                            +(double)3 * fh[idx_fh_F_ord4(iF,     jF,     kF - 4, ex)] -
-                            F32  * fh[idx_fh_F_ord4(iF,     jF,     kF - 3, ex)] +
-                            F168 * fh[idx_fh_F_ord4(iF,     jF,     kF - 2, ex)] -
-                            F672 * fh[idx_fh_F_ord4(iF,     jF,     kF - 1, ex)] +
-                            F672 * fh[idx_fh_F_ord4(iF,     jF,     kF + 1, ex)] -
-                            F168 * fh[idx_fh_F_ord4(iF,     jF,     kF + 2, ex)] +
-                            F32  * fh[idx_fh_F_ord4(iF,     jF,     kF + 3, ex)] -
-                            (double)3 * fh[idx_fh_F_ord4(iF,     jF,     kF + 4, ex)]);
-                    }
-                }
-            }
-        }
-        return;
     }
-#else
-#error "fderivs_c.C: unsupported ghost_width (must be 2, 3, 4, or 5)"
-#endif
+
+    // free(fh);
 }

AMSS_NCKU_source/fderivs_sh_c.C

@@ -1,234 +0,0 @@
-#include "macrodef.h"
-#include "share_func.h"
-
-/*
- * C 版 fderivs_sh — first derivatives on shell patch in (rho, sigma, R) coords.
- *
- * Same stencil coefficients as Cartesian fderivs, but:
- * - Uses symmetry_stbd (ghost on BOTH sides of x/y, none in z)
- * - fh buffer: (-ord+1:ex+ord) in x/y, (1:ex) in z
- * - SoA is 2-element only (x/y), no z-symmetry
- * - sst parameter (shell surface type, not used in stencil computation)
- */
-extern "C" void fderivs_sh_(const int ex[3],
-                const double *f,
-                double *fx, double *fy, double *fz,
-                const double *X, const double *Y, const double *Z,
-                double SYM1, double SYM2, double SYM3,
-                int Symmetry, int onoff, int sst)
-{
-    (void)SYM3; (void)onoff; (void)sst;
-
-    const double ZEO = 0.0, ONE = 1.0, TWO = 2.0, EIT = 8.0;
-    const double F9 = 9.0, F12 = 12.0, F45 = 45.0, F60 = 60.0;
-    const double F32 = 32.0, F168 = 168.0, F672 = 672.0, F840 = 840.0;
-
-    const int NO_SYMM = 0, EQ_SYMM = 1, OCTANT = 2;
-    const int ex1 = ex[0], ex2 = ex[1], ex3 = ex[2];
-    const double dX = X[1] - X[0];
-    const double dY = Y[1] - Y[0];
-    const double dZ = Z[1] - Z[0];
-    const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
-    const double SoA[2] = { SYM1, SYM2 };
-
-#if (ghost_width == 2)
-    {
-        const int ord = 1;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry == OCTANT && fabs(X[0]) < dX) iminF = 0;
-        if (Symmetry == OCTANT && fabs(Y[0]) < dY) jminF = 0;
-        if ((sst==2||sst==4) && fabs(Y[0]) < dY) jminF = 0; // EQ reflection
-
-        const size_t nx = (size_t)ex1 + 2 * ord;
-        const size_t ny = (size_t)ex2 + 2 * ord;
-        const size_t nz = (size_t)ex3;
-        const size_t fh_size = nx * ny * nz;
-        static double *fh_buf = NULL; static size_t cap = 0;
-        if (fh_size > cap) { free(fh_buf); fh_buf = (double*)aligned_alloc(64, fh_size*sizeof(double)); cap = fh_size; }
-        double *fh = fh_buf; if (!fh) return;
-        symmetry_stbd(ord, ex, f, fh, SoA);
-
-        const double d2dx = ONE/TWO/dX, d2dy = ONE/TWO/dY, d2dz = ONE/TWO/dZ;
-        const size_t all = (size_t)ex1*ex2*ex3;
-        for (size_t p=0;p<all;++p) { fx[p]=ZEO; fy[p]=ZEO; fz[p]=ZEO; }
-
-        const int i2_lo=(iminF>0)?iminF:0, j2_lo=(jminF>0)?jminF:0, k2_lo=1;
-        const int i2_hi=ex1-2, j2_hi=ex2-2, k2_hi=ex3-2;
-        if (i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi) {
-            for (int k0=k2_lo;k0<=k2_hi;++k0) { const int kF=k0+1;
-            for (int j0=j2_lo;j0<=j2_hi;++j0) { const int jF=j0+1;
-            for (int i0=i2_lo;i0<=i2_hi;++i0) { const int iF=i0+1;
-                const size_t p=idx_ex(i0,j0,k0,ex);
-                fx[p]=d2dx*(-fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)]);
-                fy[p]=d2dy*(-fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)]);
-                fz[p]=d2dz*(-fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)]);
-            }}}
-        }
-        return;
-    }
-#elif (ghost_width == 3)
-    {
-        const int ord = 2;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry == OCTANT && fabs(X[0]) < dX) iminF = -1;
-        if (Symmetry == OCTANT && fabs(Y[0]) < dY) jminF = -1;
-        if ((sst==2||sst==4) && fabs(Y[0]) < dY) jminF = -1;
-
-        const size_t nx=(size_t)ex1+2*ord, ny=(size_t)ex2+2*ord, nz=(size_t)ex3;
-        const size_t fh_size=nx*ny*nz;
-        static double *fh_buf=NULL; static size_t cap=0;
-        if (fh_size>cap){free(fh_buf);fh_buf=(double*)aligned_alloc(64,fh_size*sizeof(double));cap=fh_size;}
-        double *fh=fh_buf; if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const double d12dx=ONE/F12/dX, d12dy=ONE/F12/dY, d12dz=ONE/F12/dZ;
-        const double d2dx=ONE/TWO/dX, d2dy=ONE/TWO/dY, d2dz=ONE/TWO/dZ;
-        const size_t all=(size_t)ex1*ex2*ex3;
-        for(size_t p=0;p<all;++p){fx[p]=ZEO;fy[p]=ZEO;fz[p]=ZEO;}
-
-        const int i2_lo=(iminF>0)?iminF:0, j2_lo=(jminF>0)?jminF:0, k2_lo=1;
-        const int i2_hi=ex1-2, j2_hi=ex2-2, k2_hi=ex3-2;
-        const int i4_lo=(iminF+1>0)?iminF+1:0, j4_lo=(jminF+1>0)?jminF+1:0, k4_lo=2;
-        const int i4_hi=ex1-3, j4_hi=ex2-3, k4_hi=ex3-3;
-
-        if (i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi) {
-            for(int k0=k2_lo;k0<=k2_hi;++k0){const int kF=k0+1;
-            for(int j0=j2_lo;j0<=j2_hi;++j0){const int jF=j0+1;
-            for(int i0=i2_lo;i0<=i2_hi;++i0){const int iF=i0+1;
-                const size_t p=idx_ex(i0,j0,k0,ex);
-                fx[p]=d2dx*(-fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)]);
-                fy[p]=d2dy*(-fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)]);
-                fz[p]=d2dz*(-fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)]);
-            }}}
-        }
-        if (i4_lo<=i4_hi&&j4_lo<=j4_hi&&k4_lo<=k4_hi) {
-            for(int k0=k4_lo;k0<=k4_hi;++k0){const int kF=k0+1;
-            for(int j0=j4_lo;j0<=j4_hi;++j0){const int jF=j0+1;
-            for(int i0=i4_lo;i0<=i4_hi;++i0){const int iF=i0+1;
-                const size_t p=idx_ex(i0,j0,k0,ex);
-                fx[p]=d12dx*(fh[idx_fh_stbd(iF-2,jF,kF,ord,ex)]-EIT*fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+EIT*fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)]-fh[idx_fh_stbd(iF+2,jF,kF,ord,ex)]);
-                fy[p]=d12dy*(fh[idx_fh_stbd(iF,jF-2,kF,ord,ex)]-EIT*fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+EIT*fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)]-fh[idx_fh_stbd(iF,jF+2,kF,ord,ex)]);
-                fz[p]=d12dz*(fh[idx_fh_stbd(iF,jF,kF-2,ord,ex)]-EIT*fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+EIT*fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)]-fh[idx_fh_stbd(iF,jF,kF+2,ord,ex)]);
-            }}}
-        }
-        return;
-    }
-#elif (ghost_width == 4)
-    {
-        const int ord = 3;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-2;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-2;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-2;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3;
-        const size_t fh_size=nx*ny*nz;
-        static double *fh_buf=NULL;static size_t cap=0;
-        if(fh_size>cap){free(fh_buf);fh_buf=(double*)aligned_alloc(64,fh_size*sizeof(double));cap=fh_size;}
-        double *fh=fh_buf;if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const double d60dx=ONE/F60/dX,d60dy=ONE/F60/dY,d60dz=ONE/F60/dZ;
-        const double d12dx=ONE/F12/dX,d12dy=ONE/F12/dY,d12dz=ONE/F12/dZ;
-        const double d2dx=ONE/TWO/dX,d2dy=ONE/TWO/dY,d2dz=ONE/TWO/dZ;
-        const size_t all=(size_t)ex1*ex2*ex3;
-        for(size_t p=0;p<all;++p){fx[p]=ZEO;fy[p]=ZEO;fz[p]=ZEO;}
-
-        const int i2_lo=(iminF>0)?iminF:0,j2_lo=(jminF>0)?jminF:0,k2_lo=1,i2_hi=ex1-2,j2_hi=ex2-2,k2_hi=ex3-2;
-        const int i4_lo=(iminF+1>0)?iminF+1:0,j4_lo=(jminF+1>0)?jminF+1:0,k4_lo=2,i4_hi=ex1-3,j4_hi=ex2-3,k4_hi=ex3-3;
-        const int i6_lo=(iminF+2>0)?iminF+2:0,j6_lo=(jminF+2>0)?jminF+2:0,k6_lo=3,i6_hi=ex1-4,j6_hi=ex2-4,k6_hi=ex3-4;
-
-        if(i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi){
-            for(int k0=k2_lo;k0<=k2_hi;++k0){const int kF=k0+1;
-            for(int j0=j2_lo;j0<=j2_hi;++j0){const int jF=j0+1;
-            for(int i0=i2_lo;i0<=i2_hi;++i0){const int iF=i0+1;
-                const size_t p=idx_ex(i0,j0,k0,ex);
-                fx[p]=d2dx*(-fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)]);
-                fy[p]=d2dy*(-fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)]);
-                fz[p]=d2dz*(-fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)]);
-            }}}
-        }
-        if(i4_lo<=i4_hi&&j4_lo<=j4_hi&&k4_lo<=k4_hi){
-            for(int k0=k4_lo;k0<=k4_hi;++k0){const int kF=k0+1;
-            for(int j0=j4_lo;j0<=j4_hi;++j0){const int jF=j0+1;
-            for(int i0=i4_lo;i0<=i4_hi;++i0){const int iF=i0+1;
-                const size_t p=idx_ex(i0,j0,k0,ex);
-                fx[p]=d12dx*(fh[idx_fh_stbd(iF-2,jF,kF,ord,ex)]-EIT*fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+EIT*fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)]-fh[idx_fh_stbd(iF+2,jF,kF,ord,ex)]);
-                fy[p]=d12dy*(fh[idx_fh_stbd(iF,jF-2,kF,ord,ex)]-EIT*fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+EIT*fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)]-fh[idx_fh_stbd(iF,jF+2,kF,ord,ex)]);
-                fz[p]=d12dz*(fh[idx_fh_stbd(iF,jF,kF-2,ord,ex)]-EIT*fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+EIT*fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)]-fh[idx_fh_stbd(iF,jF,kF+2,ord,ex)]);
-            }}}
-        }
-        if(i6_lo<=i6_hi&&j6_lo<=j6_hi&&k6_lo<=k6_hi){
-            for(int k0=k6_lo;k0<=k6_hi;++k0){const int kF=k0+1;
-            for(int j0=j6_lo;j0<=j6_hi;++j0){const int jF=j0+1;
-            for(int i0=i6_lo;i0<=i6_hi;++i0){const int iF=i0+1;
-                const size_t p=idx_ex(i0,j0,k0,ex);
-                fx[p]=d60dx*(-fh[idx_fh_stbd(iF-3,jF,kF,ord,ex)]+F9*fh[idx_fh_stbd(iF-2,jF,kF,ord,ex)]-F45*fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+F45*fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)]-F9*fh[idx_fh_stbd(iF+2,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+3,jF,kF,ord,ex)]);
-                fy[p]=d60dy*(-fh[idx_fh_stbd(iF,jF-3,kF,ord,ex)]+F9*fh[idx_fh_stbd(iF,jF-2,kF,ord,ex)]-F45*fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+F45*fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)]-F9*fh[idx_fh_stbd(iF,jF+2,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+3,kF,ord,ex)]);
-                fz[p]=d60dz*(-fh[idx_fh_stbd(iF,jF,kF-3,ord,ex)]+F9*fh[idx_fh_stbd(iF,jF,kF-2,ord,ex)]-F45*fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+F45*fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)]-F9*fh[idx_fh_stbd(iF,jF,kF+2,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+3,ord,ex)]);
-            }}}
-        }
-        return;
-    }
-#elif (ghost_width == 5)
-    {
-        const int ord = 4;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-3;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-3;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-3;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3;
-        const size_t fh_size=nx*ny*nz;
-        static double *fh_buf=NULL;static size_t cap=0;
-        if(fh_size>cap){free(fh_buf);fh_buf=(double*)aligned_alloc(64,fh_size*sizeof(double));cap=fh_size;}
-        double *fh=fh_buf;if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const double d840dx=ONE/F840/dX,d840dy=ONE/F840/dY,d840dz=ONE/F840/dZ;
-        const double d60dx=ONE/F60/dX,d60dy=ONE/F60/dY,d60dz=ONE/F60/dZ;
-        const double d12dx=ONE/F12/dX,d12dy=ONE/F12/dY,d12dz=ONE/F12/dZ;
-        const double d2dx=ONE/TWO/dX,d2dy=ONE/TWO/dY,d2dz=ONE/TWO/dZ;
-        const size_t all=(size_t)ex1*ex2*ex3;
-        for(size_t p=0;p<all;++p){fx[p]=ZEO;fy[p]=ZEO;fz[p]=ZEO;}
-
-        const int i2_lo=(iminF>0)?iminF:0,j2_lo=(jminF>0)?jminF:0,k2_lo=1,i2_hi=ex1-2,j2_hi=ex2-2,k2_hi=ex3-2;
-        const int i4_lo=(iminF+1>0)?iminF+1:0,j4_lo=(jminF+1>0)?jminF+1:0,k4_lo=2,i4_hi=ex1-3,j4_hi=ex2-3,k4_hi=ex3-3;
-        const int i6_lo=(iminF+2>0)?iminF+2:0,j6_lo=(jminF+2>0)?jminF+2:0,k6_lo=3,i6_hi=ex1-4,j6_hi=ex2-4,k6_hi=ex3-4;
-        const int i8_lo=(iminF+3>0)?iminF+3:0,j8_lo=(jminF+3>0)?jminF+3:0,k8_lo=4,i8_hi=ex1-5,j8_hi=ex2-5,k8_hi=ex3-5;
-
-        #define FH_S(iF,jF,kF) fh[idx_fh_stbd(iF,jF,kF,ord,ex)]
-        if(i2_lo<=i2_hi&&j2_lo<=j2_hi&&k2_lo<=k2_hi){for(int k0=k2_lo;k0<=k2_hi;++k0){const int kF=k0+1;
-        for(int j0=j2_lo;j0<=j2_hi;++j0){const int jF=j0+1;
-        for(int i0=i2_lo;i0<=i2_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fx[p]=d2dx*(-FH_S(iF-1,jF,kF)+FH_S(iF+1,jF,kF));
-            fy[p]=d2dy*(-FH_S(iF,jF-1,kF)+FH_S(iF,jF+1,kF));
-            fz[p]=d2dz*(-FH_S(iF,jF,kF-1)+FH_S(iF,jF,kF+1));}}}}
-
-        if(i4_lo<=i4_hi&&j4_lo<=j4_hi&&k4_lo<=k4_hi){for(int k0=k4_lo;k0<=k4_hi;++k0){const int kF=k0+1;
-        for(int j0=j4_lo;j0<=j4_hi;++j0){const int jF=j0+1;
-        for(int i0=i4_lo;i0<=i4_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fx[p]=d12dx*(FH_S(iF-2,jF,kF)-EIT*FH_S(iF-1,jF,kF)+EIT*FH_S(iF+1,jF,kF)-FH_S(iF+2,jF,kF));
-            fy[p]=d12dy*(FH_S(iF,jF-2,kF)-EIT*FH_S(iF,jF-1,kF)+EIT*FH_S(iF,jF+1,kF)-FH_S(iF,jF+2,kF));
-            fz[p]=d12dz*(FH_S(iF,jF,kF-2)-EIT*FH_S(iF,jF,kF-1)+EIT*FH_S(iF,jF,kF+1)-FH_S(iF,jF,kF+2));}}}}
-
-        if(i6_lo<=i6_hi&&j6_lo<=j6_hi&&k6_lo<=k6_hi){for(int k0=k6_lo;k0<=k6_hi;++k0){const int kF=k0+1;
-        for(int j0=j6_lo;j0<=j6_hi;++j0){const int jF=j0+1;
-        for(int i0=i6_lo;i0<=i6_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fx[p]=d60dx*(-FH_S(iF-3,jF,kF)+F9*FH_S(iF-2,jF,kF)-F45*FH_S(iF-1,jF,kF)+F45*FH_S(iF+1,jF,kF)-F9*FH_S(iF+2,jF,kF)+FH_S(iF+3,jF,kF));
-            fy[p]=d60dy*(-FH_S(iF,jF-3,kF)+F9*FH_S(iF,jF-2,kF)-F45*FH_S(iF,jF-1,kF)+F45*FH_S(iF,jF+1,kF)-F9*FH_S(iF,jF+2,kF)+FH_S(iF,jF+3,kF));
-            fz[p]=d60dz*(-FH_S(iF,jF,kF-3)+F9*FH_S(iF,jF,kF-2)-F45*FH_S(iF,jF,kF-1)+F45*FH_S(iF,jF,kF+1)-F9*FH_S(iF,jF,kF+2)+FH_S(iF,jF,kF+3));}}}}
-
-        if(i8_lo<=i8_hi&&j8_lo<=j8_hi&&k8_lo<=k8_hi){for(int k0=k8_lo;k0<=k8_hi;++k0){const int kF=k0+1;
-        for(int j0=j8_lo;j0<=j8_hi;++j0){const int jF=j0+1;
-        for(int i0=i8_lo;i0<=i8_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-            fx[p]=d840dx*(+(double)3*FH_S(iF-4,jF,kF)-F32*FH_S(iF-3,jF,kF)+F168*FH_S(iF-2,jF,kF)-F672*FH_S(iF-1,jF,kF)+F672*FH_S(iF+1,jF,kF)-F168*FH_S(iF+2,jF,kF)+F32*FH_S(iF+3,jF,kF)-(double)3*FH_S(iF+4,jF,kF));
-            fy[p]=d840dy*(+(double)3*FH_S(iF,jF-4,kF)-F32*FH_S(iF,jF-3,kF)+F168*FH_S(iF,jF-2,kF)-F672*FH_S(iF,jF-1,kF)+F672*FH_S(iF,jF+1,kF)-F168*FH_S(iF,jF+2,kF)+F32*FH_S(iF,jF+3,kF)-(double)3*FH_S(iF,jF+4,kF));
-            fz[p]=d840dz*(+(double)3*FH_S(iF,jF,kF-4)-F32*FH_S(iF,jF,kF-3)+F168*FH_S(iF,jF,kF-2)-F672*FH_S(iF,jF,kF-1)+F672*FH_S(iF,jF,kF+1)-F168*FH_S(iF,jF,kF+2)+F32*FH_S(iF,jF,kF+3)-(double)3*FH_S(iF,jF,kF+4));}}}}
-        #undef FH_S
-        return;
-    }
-#else
-#error "fderivs_sh_c.C: unsupported ghost_width"
-#endif
-}

AMSS_NCKU_source/fderivs_shc_c.C

@@ -1,54 +0,0 @@
-#include "macrodef.h"
-#include "share_func.h"
-#include <cstddef>
-
-/*
- * fderivs_shc — shell first derivatives converted to Cartesian via chain rule.
- *
- * Calls fderivs_sh internally, then:
- *   fx = drhodx * df/drho + dsigmadx * df/dsigma + dRdx * df/dR
- *   fy = drhody * df/drho + dsigmady * df/dsigma + dRdy * df/dR
- *   fz = drhodz * df/drho + dsigmadz * df/dsigma + dRdz * df/dR
- */
-
-// Forward declaration (defined in fderivs_sh_c.C with extern "C" name fderivs_sh_)
-extern "C" {
-void fderivs_sh_(const int ex[3], const double *f,
-                 double *fx, double *fy, double *fz,
-                 const double *X, const double *Y, const double *Z,
-                 double SYM1, double SYM2, double SYM3,
-                 int Symmetry, int onoff, int sst);
-
-void fderivs_shc_(int *ex,
-                   double *f,
-                   double *fx, double *fy, double *fz,
-                   double *crho, double *sigma, double *R,
-                   double &SYM1, double &SYM2, double &SYM3,
-                   int &Symmetry, int &Lev, int &sst,
-                   double *drhodx, double *drhody, double *drhodz,
-                   double *dsigmadx, double *dsigmady, double *dsigmadz,
-                   double *dRdx, double *dRdy, double *dRdz)
-{
-    const int ex3[3] = { ex[0], ex[1], ex[2] };
-    const size_t n = (size_t)ex[0] * (size_t)ex[1] * (size_t)ex[2];
-
-    // Temporary shell-coordinate derivatives
-    double *gx = (double*)malloc(n * sizeof(double));
-    double *gy = (double*)malloc(n * sizeof(double));
-    double *gz = (double*)malloc(n * sizeof(double));
-    if (!gx || !gy || !gz) { free(gx); free(gy); free(gz); return; }
-
-    // Compute shell-coordinate derivatives
-    fderivs_sh_(ex3, f, gx, gy, gz, crho, sigma, R, SYM1, SYM2, SYM3, Symmetry, Lev, sst);
-
-    // Chain rule to Cartesian
-    for (size_t i = 0; i < n; ++i) {
-        fx[i] = drhodx[i]   * gx[i] + dsigmadx[i] * gy[i] + dRdx[i] * gz[i];
-        fy[i] = drhody[i]   * gx[i] + dsigmady[i] * gy[i] + dRdy[i] * gz[i];
-        fz[i] = drhodz[i]   * gx[i] + dsigmadz[i] * gy[i] + dRdz[i] * gz[i];
-    }
-
-    free(gx); free(gy); free(gz);
-}
-
-} // extern "C"

AMSS_NCKU_source/fmisc.f90

@@ -312,8 +312,8 @@
   end function decide3d
 
 !---------------------------------------------------------------------------------------
-subroutine symmetry_bd(ord,extc,func,funcc,SoA)
-  implicit none
+subroutine symmetry_bd(ord,extc,func,funcc,SoA)
+  implicit none
 
 !~~~~~~> input arguments
   integer,intent(in) :: ord
@@ -322,12 +322,9 @@ subroutine symmetry_bd(ord,extc,func,funcc,SoA)
   real*8, dimension(-ord+1:extc(1),-ord+1:extc(2),-ord+1:extc(3)),intent(out):: funcc
   real*8, dimension(1:3), intent(in) :: SoA
 
-  integer::i
-
-#if USE_FMISC_SAFE_MODE
-  funcc = 0.d0
-#endif
-  funcc(1:extc(1),1:extc(2),1:extc(3)) = func
+  integer::i
+
+  funcc(1:extc(1),1:extc(2),1:extc(3)) = func
    do i=0,ord-1
       funcc(-i,1:extc(2),1:extc(3)) = funcc(i+2,1:extc(2),1:extc(3))*SoA(1)
    enddo
@@ -340,8 +337,8 @@ subroutine symmetry_bd(ord,extc,func,funcc,SoA)
 
 end subroutine symmetry_bd
 
-subroutine symmetry_tbd(ord,extc,func,funcc,SoA)
-  implicit none
+subroutine symmetry_tbd(ord,extc,func,funcc,SoA)
+  implicit none
 
 !~~~~~~> input arguments
   integer,intent(in) :: ord
@@ -350,12 +347,9 @@ subroutine symmetry_tbd(ord,extc,func,funcc,SoA)
   real*8, dimension(-ord+1:extc(1)+ord,-ord+1:extc(2)+ord,-ord+1:extc(3)+ord),intent(out):: funcc
   real*8, dimension(1:3), intent(in) :: SoA
 
-  integer::i
-
-#if USE_FMISC_SAFE_MODE
-  funcc = 0.d0
-#endif
-  funcc(1:extc(1),1:extc(2),1:extc(3)) = func
+  integer::i
+
+  funcc(1:extc(1),1:extc(2),1:extc(3)) = func
    do i=0,ord-1
       funcc(-i,1:extc(2),1:extc(3)) = funcc(i+2,1:extc(2),1:extc(3))*SoA(1)
       funcc(extc(1)+1+i,1:extc(2),1:extc(3)) = funcc(extc(1)-1-i,1:extc(2),1:extc(3))*SoA(1)
@@ -371,8 +365,8 @@ subroutine symmetry_tbd(ord,extc,func,funcc,SoA)
 
 end subroutine symmetry_tbd
 
-subroutine symmetry_stbd(ord,extc,func,funcc,SoA)
-  implicit none
+subroutine symmetry_stbd(ord,extc,func,funcc,SoA)
+  implicit none
 
 !~~~~~~> input arguments
   integer,intent(in) :: ord
@@ -381,12 +375,9 @@ subroutine symmetry_stbd(ord,extc,func,funcc,SoA)
   real*8, dimension(-ord+1:extc(1)+ord,-ord+1:extc(2)+ord,extc(3)),intent(out):: funcc
   real*8, dimension(2), intent(in) :: SoA
 
-  integer::i
-
-#if USE_FMISC_SAFE_MODE
-  funcc = 0.d0
-#endif
-  funcc(1:extc(1),1:extc(2),1:extc(3)) = func
+  integer::i
+
+  funcc(1:extc(1),1:extc(2),1:extc(3)) = func
    do i=0,ord-1
       funcc(-i,1:extc(2),1:extc(3)) = funcc(i+2,1:extc(2),1:extc(3))*SoA(1)
       funcc(extc(1)+1+i,1:extc(2),1:extc(3)) = funcc(extc(1)-1-i,1:extc(2),1:extc(3))*SoA(1)
@@ -1128,10 +1119,6 @@ end subroutine d2dump
 #define POLINT6_USE_BARYCENTRIC 1
 #endif
 
-#ifndef USE_FMISC_SAFE_MODE
-#define USE_FMISC_SAFE_MODE 0
-#endif
-
 !DIR$ ATTRIBUTES FORCEINLINE :: polint6_neville
   subroutine polint6_neville(xa, ya, x, y, dy)
   implicit none
@@ -1284,9 +1271,7 @@ end subroutine d2dump
   real*8 :: dif, dift, hp, h, den_val
 
   if (ordn == 6) then
-#if USE_FMISC_SAFE_MODE
-    call polint6_neville(xa, ya, x, y, dy)
-#elif POLINT6_USE_BARYCENTRIC
+#if POLINT6_USE_BARYCENTRIC
     call polint6_barycentric(xa, ya, x, y, dy)
 #else
     call polint6_neville(xa, ya, x, y, dy)
@@ -1391,10 +1376,10 @@ end subroutine d2dump
   real*8, intent(in) :: x1,x2
   real*8, intent(out) :: y,dy
 
-#if USE_FMISC_SAFE_MODE || defined(POLINT_LEGACY_ORDER)
-  integer  :: i,m
-  real*8, dimension(ordn) :: ymtmp
-  real*8, dimension(ordn) :: yntmp
+#ifdef POLINT_LEGACY_ORDER
+  integer  :: i,m
+  real*8, dimension(ordn) :: ymtmp
+  real*8, dimension(ordn) :: yntmp
 
   m=size(x1a)
   do i=1,m
@@ -1429,7 +1414,7 @@ end subroutine d2dump
   real*8, intent(in) :: x1,x2,x3
   real*8, intent(out) :: y,dy
 
-#if USE_FMISC_SAFE_MODE || defined(POLINT_LEGACY_ORDER)
+#ifdef POLINT_LEGACY_ORDER
   integer  :: i,j,m,n
   real*8, dimension(ordn,ordn) :: yatmp
   real*8, dimension(ordn) :: ymtmp
@@ -1517,110 +1502,22 @@ if(dabs(X(1)-xmin) < dX) imin = 1
 if(dabs(Y(1)-ymin) < dY) jmin = 1
 if(dabs(Z(1)-zmin) < dZ) kmin = 1
 
-#if USE_FMISC_SAFE_MODE
-  f_out = 0.d0
-  do k = kmin, kmax
-    do j = jmin, jmax
-      do i = imin, imax
-        f_out = f_out + f(i,j,k)*f(i,j,k)
-      end do
-    end do
-  end do
-#else
-! Optimized with oneMKL BLAS DDOT for dot product
-  n_elements = (imax-imin+1)*(jmax-jmin+1)*(kmax-kmin+1)
-  allocate(f_flat(n_elements))
-  f_flat = reshape(f(imin:imax,jmin:jmax,kmin:kmax), [n_elements])
-  f_out = DDOT(n_elements, f_flat, 1, f_flat, 1)
-  deallocate(f_flat)
-#endif
+! Optimized with oneMKL BLAS DDOT for dot product
+n_elements = (imax-imin+1)*(jmax-jmin+1)*(kmax-kmin+1)
+allocate(f_flat(n_elements))
+f_flat = reshape(f(imin:imax,jmin:jmax,kmin:kmax), [n_elements])
+f_out = DDOT(n_elements, f_flat, 1, f_flat, 1)
+deallocate(f_flat)
 
 f_out = f_out*dX*dY*dZ
 
-  return
-
-  end subroutine l2normhelper
-!--------------------------------------------------------------------------------------
-  subroutine l2normhelper7(ex, X, Y, Z,xmin,ymin,zmin,xmax,ymax,zmax,&
-                           f1,f2,f3,f4,f5,f6,f7,f_out,gw)
-
-  implicit none
-!~~~~~~> Input parameters:
-  integer,intent(in ):: ex(1:3)
-  real*8, intent(in ):: X(1:ex(1)),Y(1:ex(2)),Z(1:ex(3)),xmin,ymin,zmin,xmax,ymax,zmax
-  integer,intent(in)::gw
-  real*8, dimension(ex(1),ex(2),ex(3)),intent(in) :: f1,f2,f3,f4,f5,f6,f7
-  real*8, intent(out) :: f_out(7)
-!~~~~~~> Other variables:
-
-  real*8            :: dX, dY, dZ
-  integer::imin,jmin,kmin
-  integer::imax,jmax,kmax
-  integer::i,j,k
-  real*8 :: s1,s2,s3,s4,s5,s6,s7
-
-  dX = X(2) - X(1)
-  dY = Y(2) - Y(1)
-  dZ = Z(2) - Z(1)
-
-! for ghost zone
-   imin = gw+1
-   jmin = gw+1
-   kmin = gw+1
-
-   imax = ex(1) - gw
-   jmax = ex(2) - gw
-   kmax = ex(3) - gw
-
-!for patch boundary (i.e., not ghost boundary)
-
-if(dabs(X(ex(1))-xmax) < dX) imax = ex(1)
-if(dabs(Y(ex(2))-ymax) < dY) jmax = ex(2)
-if(dabs(Z(ex(3))-zmax) < dZ) kmax = ex(3)
-if(dabs(X(1)-xmin) < dX) imin = 1
-if(dabs(Y(1)-ymin) < dY) jmin = 1
-if(dabs(Z(1)-zmin) < dZ) kmin = 1
-
-  s1 = 0.d0
-  s2 = 0.d0
-  s3 = 0.d0
-  s4 = 0.d0
-  s5 = 0.d0
-  s6 = 0.d0
-  s7 = 0.d0
-
-  do k=kmin,kmax
-    do j=jmin,jmax
-#if !USE_FMISC_SAFE_MODE
-!DIR$ SIMD REDUCTION(+:s1,s2,s3,s4,s5,s6,s7)
-#endif
-      do i=imin,imax
-        s1 = s1 + f1(i,j,k)*f1(i,j,k)
-        s2 = s2 + f2(i,j,k)*f2(i,j,k)
-        s3 = s3 + f3(i,j,k)*f3(i,j,k)
-        s4 = s4 + f4(i,j,k)*f4(i,j,k)
-        s5 = s5 + f5(i,j,k)*f5(i,j,k)
-        s6 = s6 + f6(i,j,k)*f6(i,j,k)
-        s7 = s7 + f7(i,j,k)*f7(i,j,k)
-      enddo
-    enddo
-  enddo
-
-  f_out(1) = s1*dX*dY*dZ
-  f_out(2) = s2*dX*dY*dZ
-  f_out(3) = s3*dX*dY*dZ
-  f_out(4) = s4*dX*dY*dZ
-  f_out(5) = s5*dX*dY*dZ
-  f_out(6) = s6*dX*dY*dZ
-  f_out(7) = s7*dX*dY*dZ
-
-  return
-
-  end subroutine l2normhelper7
-!--------------------------------------------------------------------------------------
-! calculate L2norm especially for shell Blocks
-  subroutine l2normhelper_sh(ex, X, Y, Z,xmin,ymin,zmin,xmax,ymax,zmax,&
-                          f,f_out,gw,ogw,Symmetry)
+  return
+
+  end subroutine l2normhelper
+!--------------------------------------------------------------------------------------
+! calculate L2norm especially for shell Blocks
+  subroutine l2normhelper_sh(ex, X, Y, Z,xmin,ymin,zmin,xmax,ymax,zmax,&
+                          f,f_out,gw,ogw,Symmetry)
 
   implicit none
 !~~~~~~> Input parameters:
@@ -1700,23 +1597,12 @@ if(Symmetry==2)then
   if(dabs(ymin+gw*dY)<dY.and.Y(1)<0.d0) jmin = gw+1
 endif
 
-#if USE_FMISC_SAFE_MODE
-  f_out = 0.d0
-  do k = kmin, kmax
-    do j = jmin, jmax
-      do i = imin, imax
-        f_out = f_out + f(i,j,k)*f(i,j,k)
-      end do
-    end do
-  end do
-#else
-! Optimized with oneMKL BLAS DDOT for dot product
-  n_elements = (imax-imin+1)*(jmax-jmin+1)*(kmax-kmin+1)
-  allocate(f_flat(n_elements))
-  f_flat = reshape(f(imin:imax,jmin:jmax,kmin:kmax), [n_elements])
-  f_out = DDOT(n_elements, f_flat, 1, f_flat, 1)
-  deallocate(f_flat)
-#endif
+! Optimized with oneMKL BLAS DDOT for dot product
+n_elements = (imax-imin+1)*(jmax-jmin+1)*(kmax-kmin+1)
+allocate(f_flat(n_elements))
+f_flat = reshape(f(imin:imax,jmin:jmax,kmin:kmax), [n_elements])
+f_out = DDOT(n_elements, f_flat, 1, f_flat, 1)
+deallocate(f_flat)
 
 f_out = f_out*dX*dY*dZ
 
@@ -1808,23 +1694,12 @@ if(Symmetry==2)then
   if(dabs(ymin+gw*dY)<dY.and.Y(1)<0.d0) jmin = gw+1
 endif
 
-#if USE_FMISC_SAFE_MODE
-  f_out = 0.d0
-  do k = kmin, kmax
-    do j = jmin, jmax
-      do i = imin, imax
-        f_out = f_out + f(i,j,k)*f(i,j,k)
-      end do
-    end do
-  end do
-#else
-! Optimized with oneMKL BLAS DDOT for dot product
-  Nout = (imax-imin+1)*(jmax-jmin+1)*(kmax-kmin+1)
-  allocate(f_flat(Nout))
-  f_flat = reshape(f(imin:imax,jmin:jmax,kmin:kmax), [Nout])
-  f_out = DDOT(Nout, f_flat, 1, f_flat, 1)
-  deallocate(f_flat)
-#endif
+! Optimized with oneMKL BLAS DDOT for dot product
+Nout = (imax-imin+1)*(jmax-jmin+1)*(kmax-kmin+1)
+allocate(f_flat(Nout))
+f_flat = reshape(f(imin:imax,jmin:jmax,kmin:kmax), [Nout])
+f_out = DDOT(Nout, f_flat, 1, f_flat, 1)
+deallocate(f_flat)
 
   return
 
@@ -1928,23 +1803,13 @@ endif
   real*8,parameter::C1=3.d0/8.d0,C2=3.d0/4.d0,C3=-1.d0/8.d0
   integer :: i,j,k
 
-#if USE_FMISC_SAFE_MODE
-  do k=1,ext(3)
-   do j=1,ext(2)
-    do i=1,ext(1)
-      fout(i,j,k) = C1*f1(i,j,k)+C2*f2(i,j,k)+C3*f3(i,j,k)
-    enddo
-   enddo
-  enddo
-#else
   do concurrent (k=1:ext(3), j=1:ext(2), i=1:ext(1))
     fout(i,j,k) = C1*f1(i,j,k)+C2*f2(i,j,k)+C3*f3(i,j,k)
   end do
-#endif
 
   return
-
-  end subroutine average2
+
+  end subroutine average2
 !-----------------------------------------------------------------------------  
   subroutine average2p(ext,f1,f2,f3,fout)
   implicit none
@@ -2084,15 +1949,8 @@ endif
     tmp1 = tmp1 + coef(ORDN+m)*tmp2(:,m)
   enddo
 
-#if USE_FMISC_SAFE_MODE
-  f_int = 0.d0
-  do m = 1, ORDN
-    f_int = f_int + coef(m) * tmp1(m)
-  end do
-#else
-  ! Third dimension: x-direction weighted sum using BLAS DDOT
-  f_int = DDOT(ORDN, coef(1:ORDN), 1, tmp1, 1)
-#endif
+  ! Third dimension: x-direction weighted sum using BLAS DDOT
+  f_int = DDOT(ORDN, coef(1:ORDN), 1, tmp1, 1)
 
   return
 
@@ -2158,15 +2016,8 @@ endif
     tmp1 = tmp1 + coef(ORDN+m)*ya(:,m)
   enddo
 
-#if USE_FMISC_SAFE_MODE
-  f_int = 0.d0
-  do m = 1, ORDN
-    f_int = f_int + coef(m) * tmp1(m)
-  end do
-#else
-  ! Use BLAS DDOT for final weighted sum
-  f_int = DDOT(ORDN, coef(1:ORDN), 1, tmp1, 1)
-#endif
+  ! Use BLAS DDOT for final weighted sum
+  f_int = DDOT(ORDN, coef(1:ORDN), 1, tmp1, 1)
 
   return
 
@@ -2258,15 +2109,8 @@ endif
           write(*,*)"error in global_interpind1d, not recognized dumyd = ",dumyd
   endif
 
-#if USE_FMISC_SAFE_MODE
-  f_int = 0.d0
-  do m = 1, ORDN
-    f_int = f_int + coef(m) * ya(m)
-  end do
-#else
-  ! Optimized with BLAS DDOT for weighted sum
-  f_int = DDOT(ORDN, coef, 1, ya, 1)
-#endif
+  ! Optimized with BLAS DDOT for weighted sum
+  f_int = DDOT(ORDN, coef, 1, ya, 1)
 
   return
 

AMSS_NCKU_source/fmisc.h

@@ -12,10 +12,9 @@
 #define f_global_interpind global_interpind
 #define f_global_interpind2d global_interpind2d
 #define f_global_interpind1d global_interpind1d
-#define f_l2normhelper l2normhelper
-#define f_l2normhelper7 l2normhelper7
-#define f_l2normhelper_sh l2normhelper_sh
-#define f_l2normhelper_sh_rms l2normhelper_sh_rms
+#define f_l2normhelper l2normhelper
+#define f_l2normhelper_sh l2normhelper_sh
+#define f_l2normhelper_sh_rms l2normhelper_sh_rms
 #define f_average average
 #define f_average3 average3
 #define f_average2 average2
@@ -42,10 +41,9 @@
 #define f_global_interpind GLOBAL_INTERPIND
 #define f_global_interpind2d GLOBAL_INTERPIND2D
 #define f_global_interpind1d GLOBAL_INTERPIND1D
-#define f_l2normhelper L2NORMHELPER
-#define f_l2normhelper7 L2NORMHELPER7
-#define f_l2normhelper_sh L2NORMHELPER_SH
-#define f_l2normhelper_sh_rms L2NORMHELPER_SH_RMS
+#define f_l2normhelper L2NORMHELPER
+#define f_l2normhelper_sh L2NORMHELPER_SH
+#define f_l2normhelper_sh_rms L2NORMHELPER_SH_RMS
 #define f_average AVERAGE
 #define f_average3 AVERAGE3
 #define f_average2 AVERAGE2
@@ -72,10 +70,9 @@
 #define f_global_interpind global_interpind_
 #define f_global_interpind2d global_interpind2d_
 #define f_global_interpind1d global_interpind1d_
-#define f_l2normhelper l2normhelper_
-#define f_l2normhelper7 l2normhelper7_
-#define f_l2normhelper_sh l2normhelper_sh_
-#define f_l2normhelper_sh_rms l2normhelper_sh_rms_
+#define f_l2normhelper l2normhelper_
+#define f_l2normhelper_sh l2normhelper_sh_
+#define f_l2normhelper_sh_rms l2normhelper_sh_rms_
 #define f_average average_
 #define f_average3 average3_
 #define f_average2 average2_
@@ -159,29 +156,20 @@ extern "C"
 							  int *, double *, int &, int &);
 }
 
-extern "C"
-{
-	void f_l2normhelper(int *, double *, double *, double *,
-						double &, double &, double &,
-						double &, double &, double &,
-						double *, double &, int &);
-}
-
-extern "C"
-{
-	void f_l2normhelper7(int *, double *, double *, double *,
-						 double &, double &, double &,
-						 double &, double &, double &,
-						 double *, double *, double *, double *,
-						 double *, double *, double *, double *, int &);
-}
-
-extern "C"
-{
-	void f_l2normhelper_sh(int *, double *, double *, double *,
-						   double &, double &, double &,
-						   double &, double &, double &,
-						   double *, double &, int &, int &, int &);
+extern "C"
+{
+	void f_l2normhelper(int *, double *, double *, double *,
+						double &, double &, double &,
+						double &, double &, double &,
+						double *, double &, int &);
+}
+
+extern "C"
+{
+	void f_l2normhelper_sh(int *, double *, double *, double *,
+						   double &, double &, double &,
+						   double &, double &, double &,
+						   double *, double &, int &, int &, int &);
 }
 
 extern "C"

AMSS_NCKU_source/kodiss_c.C

@@ -1,16 +1,16 @@
-#include "macrodef.h"
 #include "tool.h"
 
 /*
- * C 版 kodis — Kreiss-Oliger numerical dissipation (Cartesian patches).
+ * C 版 kodis
  *
- * The KO operator is (D₊D₋)^r applied to f_rhs with alternating sign (-1)^(r-1).
+ * Fortran signature:
+ * subroutine kodis(ex,X,Y,Z,f,f_rhs,SoA,Symmetry,eps)
  *
- * FD order → r → cof=2^(2r) mapping:
- *   ghost_width=2 (2nd) → r=2, cof=16,  sign=-
- *   ghost_width=3 (4th) → r=3, cof=64,  sign=+
- *   ghost_width=4 (6th) → r=4, cof=256, sign=-
- *   ghost_width=5 (8th) → r=5, cof=1024,sign=+
+ * 约定：
+ *   X: ex1, Y: ex2, Z: ex3
+ *   f, f_rhs: ex1*ex2*ex3 按 idx_ex 布局
+ *   SoA[3]
+ *   eps: double
  */
 void kodis(const int ex[3],
            const double *X, const double *Y, const double *Z,
@@ -18,304 +18,100 @@ void kodis(const int ex[3],
            const double SoA[3],
            int Symmetry, double eps)
 {
-    const double ZEO = 0.0;
+    const double ONE = 1.0, SIX = 6.0, FIT = 15.0, TWT = 20.0;
+    const double cof = 64.0;             // 2^6
+    const int NO_SYMM = 0, OCTANT = 2;
+
     const int ex1 = ex[0], ex2 = ex[1], ex3 = ex[2];
+
+    // Fortran: dX = X(2)-X(1) -> C: X[1]-X[0]
     const double dX = X[1] - X[0];
     const double dY = Y[1] - Y[0];
     const double dZ = Z[1] - Z[0];
+    (void)ONE; // ONE 在原 Fortran 里只是参数，这里不一定用得上
 
-    const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
+    // Fortran: imax=ex(1) 等是 1-based 上界
+    const int imaxF = ex1;
+    const int jmaxF = ex2;
+    const int kmaxF = ex3;
 
-#if (ghost_width == 2)
-    /* ---- r=2, cof=16, sign=-, 5pt stencil ----------------------------- */
-    {
-        const int ord = 2;
-        const int r = 2;
-        const double cof = 16.0;
-        const double F4 = 4.0, F6 = 6.0;
-        const int NO_SYMM = 0, EQ_SYMM = 1;
+    // Fortran: imin=jmin=kmin=1，某些对称情况变 -2
+    int iminF = 1, jminF = 1, kminF = 1;
 
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -1;
+    if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -2;
+    if (Symmetry == OCTANT && fabs(X[0]) < dX) iminF = -2;
+    if (Symmetry == OCTANT && fabs(Y[0]) < dY) jminF = -2;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
+    // 分配 fh：大小 (ex1+3)*(ex2+3)*(ex3+3)，对应 ord=3
+    const size_t nx = (size_t)ex1 + 3;
+    const size_t ny = (size_t)ex2 + 3;
+    const size_t nz = (size_t)ex3 + 3;
+    const size_t fh_size = nx * ny * nz;
 
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
+    double *fh = (double*)malloc(fh_size * sizeof(double));
+    if (!fh) return;
 
-        symmetry_bd(ord, ex, f, fh, SoA);
+    // Fortran: call symmetry_bd(3,ex,f,fh,SoA)
+    symmetry_bd(3, ex, f, fh, SoA);
 
-        /* i±2 must be valid: i-2 >= iminF && i+2 <= imaxF
-           C 0-based: i0 >= iminF+1, i0 <= ex1-3 */
-        const int i0_lo = (iminF + 1 > 0) ? (iminF + 1) : 0;
-        const int j0_lo = (jminF + 1 > 0) ? (jminF + 1) : 0;
-        const int k0_lo = (kminF + 1 > 0) ? (kminF + 1) : 0;
-        const int i0_hi = imaxF - 3;
-        const int j0_hi = jmaxF - 3;
-        const int k0_hi = kmaxF - 3;
+    /*
+     * Fortran loops:
+     * do k=1,ex3
+     * do j=1,ex2
+     * do i=1,ex1
+     *
+     * C: k0=0..ex3-1, j0=0..ex2-1, i0=0..ex1-1
+     * 并定义 Fortran index: iF=i0+1, ...
+     */
+    // 收紧循环范围：只遍历满足 iF±3/jF±3/kF±3 条件的内部点
+    // iF-3 >= iminF => iF >= iminF+3 => i0 >= iminF+2 (因为 iF=i0+1)
+    // iF+3 <= imaxF => iF <= imaxF-3 => i0 <= imaxF-4
+    const int i0_lo = (iminF + 2 > 0) ? iminF + 2 : 0;
+    const int j0_lo = (jminF + 2 > 0) ? jminF + 2 : 0;
+    const int k0_lo = (kminF + 2 > 0) ? kminF + 2 : 0;
+    const int i0_hi = imaxF - 4;  // inclusive
+    const int j0_hi = jmaxF - 4;
+    const int k0_hi = kmaxF - 4;
 
-        if (!(i0_lo > i0_hi || j0_lo > j0_hi || k0_lo > k0_hi)) {
-            for (int k0 = k0_lo; k0 <= k0_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j0_lo; j0 <= j0_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i0_lo; i0 <= i0_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        const double Dx = (
-                            (fh[idx_fh_F_ord2(iF - 2, jF, kF, ex)] + fh[idx_fh_F_ord2(iF + 2, jF, kF, ex)]) -
-                            F4 * (fh[idx_fh_F_ord2(iF - 1, jF, kF, ex)] + fh[idx_fh_F_ord2(iF + 1, jF, kF, ex)]) +
-                            F6 *  fh[idx_fh_F_ord2(iF,     jF, kF, ex)]
-                        ) / dX;
-
-                        const double Dy = (
-                            (fh[idx_fh_F_ord2(iF, jF - 2, kF, ex)] + fh[idx_fh_F_ord2(iF, jF + 2, kF, ex)]) -
-                            F4 * (fh[idx_fh_F_ord2(iF, jF - 1, kF, ex)] + fh[idx_fh_F_ord2(iF, jF + 1, kF, ex)]) +
-                            F6 *  fh[idx_fh_F_ord2(iF, jF,     kF, ex)]
-                        ) / dY;
-
-                        const double Dz = (
-                            (fh[idx_fh_F_ord2(iF, jF, kF - 2, ex)] + fh[idx_fh_F_ord2(iF, jF, kF + 2, ex)]) -
-                            F4 * (fh[idx_fh_F_ord2(iF, jF, kF - 1, ex)] + fh[idx_fh_F_ord2(iF, jF, kF + 1, ex)]) +
-                            F6 *  fh[idx_fh_F_ord2(iF, jF, kF,     ex)]
-                        ) / dZ;
-
-                        f_rhs[p] -= (eps / cof) * (Dx + Dy + Dz); /* sign=- */
-                    }
-                }
-            }
-        }
+    if (i0_lo > i0_hi || j0_lo > j0_hi || k0_lo > k0_hi) {
         free(fh);
         return;
     }
-#elif (ghost_width == 3)
-    /* ---- r=3, cof=64, sign=+, 7pt stencil (current default) ---------- */
-    {
-        const int ord = 3;
-        const int r = 3;
-        const double cof = 64.0;
-        const double SIX = 6.0, FIT = 15.0, TWT = 20.0;
-        const int NO_SYMM = 0, OCTANT = 2;
 
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -2;
-        if (Symmetry == OCTANT && fabs(X[0]) < dX) iminF = -2;
-        if (Symmetry == OCTANT && fabs(Y[0]) < dY) jminF = -2;
+    for (int k0 = k0_lo; k0 <= k0_hi; ++k0) {
+        const int kF = k0 + 1;
+        for (int j0 = j0_lo; j0 <= j0_hi; ++j0) {
+            const int jF = j0 + 1;
+            for (int i0 = i0_lo; i0 <= i0_hi; ++i0) {
+                const int iF = i0 + 1;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
+                    const size_t p = idx_ex(i0, j0, k0, ex);
 
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
+                    // 三个方向各一份同型的 7 点组合（实际上是对称的 6th-order dissipation/filter 核）
+                    const double Dx_term =
+                        ( (fh[idx_fh_F(iF - 3, jF, kF, ex)] + fh[idx_fh_F(iF + 3, jF, kF, ex)]) -
+                          SIX * (fh[idx_fh_F(iF - 2, jF, kF, ex)] + fh[idx_fh_F(iF + 2, jF, kF, ex)]) +
+                          FIT * (fh[idx_fh_F(iF - 1, jF, kF, ex)] + fh[idx_fh_F(iF + 1, jF, kF, ex)]) -
+                          TWT *  fh[idx_fh_F(iF    , jF, kF, ex)] ) / dX;
 
-        symmetry_bd(ord, ex, f, fh, SoA);
+                    const double Dy_term =
+                        ( (fh[idx_fh_F(iF, jF - 3, kF, ex)] + fh[idx_fh_F(iF, jF + 3, kF, ex)]) -
+                          SIX * (fh[idx_fh_F(iF, jF - 2, kF, ex)] + fh[idx_fh_F(iF, jF + 2, kF, ex)]) +
+                          FIT * (fh[idx_fh_F(iF, jF - 1, kF, ex)] + fh[idx_fh_F(iF, jF + 1, kF, ex)]) -
+                          TWT *  fh[idx_fh_F(iF, jF    , kF, ex)] ) / dY;
 
-        const int i0_lo = (iminF + 2 > 0) ? iminF + 2 : 0;
-        const int j0_lo = (jminF + 2 > 0) ? jminF + 2 : 0;
-        const int k0_lo = (kminF + 2 > 0) ? kminF + 2 : 0;
-        const int i0_hi = imaxF - 4;
-        const int j0_hi = jmaxF - 4;
-        const int k0_hi = kmaxF - 4;
+                    const double Dz_term =
+                        ( (fh[idx_fh_F(iF, jF, kF - 3, ex)] + fh[idx_fh_F(iF, jF, kF + 3, ex)]) -
+                          SIX * (fh[idx_fh_F(iF, jF, kF - 2, ex)] + fh[idx_fh_F(iF, jF, kF + 2, ex)]) +
+                          FIT * (fh[idx_fh_F(iF, jF, kF - 1, ex)] + fh[idx_fh_F(iF, jF, kF + 1, ex)]) -
+                          TWT *  fh[idx_fh_F(iF, jF, kF    , ex)] ) / dZ;
 
-        if (!(i0_lo > i0_hi || j0_lo > j0_hi || k0_lo > k0_hi)) {
-            for (int k0 = k0_lo; k0 <= k0_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j0_lo; j0 <= j0_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i0_lo; i0 <= i0_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        const double Dx = (
-                            (fh[idx_fh_F(iF - 3, jF, kF, ex)] + fh[idx_fh_F(iF + 3, jF, kF, ex)]) -
-                            SIX * (fh[idx_fh_F(iF - 2, jF, kF, ex)] + fh[idx_fh_F(iF + 2, jF, kF, ex)]) +
-                            FIT * (fh[idx_fh_F(iF - 1, jF, kF, ex)] + fh[idx_fh_F(iF + 1, jF, kF, ex)]) -
-                            TWT *  fh[idx_fh_F(iF,     jF, kF, ex)]
-                        ) / dX;
-
-                        const double Dy = (
-                            (fh[idx_fh_F(iF, jF - 3, kF, ex)] + fh[idx_fh_F(iF, jF + 3, kF, ex)]) -
-                            SIX * (fh[idx_fh_F(iF, jF - 2, kF, ex)] + fh[idx_fh_F(iF, jF + 2, kF, ex)]) +
-                            FIT * (fh[idx_fh_F(iF, jF - 1, kF, ex)] + fh[idx_fh_F(iF, jF + 1, kF, ex)]) -
-                            TWT *  fh[idx_fh_F(iF, jF,     kF, ex)]
-                        ) / dY;
-
-                        const double Dz = (
-                            (fh[idx_fh_F(iF, jF, kF - 3, ex)] + fh[idx_fh_F(iF, jF, kF + 3, ex)]) -
-                            SIX * (fh[idx_fh_F(iF, jF, kF - 2, ex)] + fh[idx_fh_F(iF, jF, kF + 2, ex)]) +
-                            FIT * (fh[idx_fh_F(iF, jF, kF - 1, ex)] + fh[idx_fh_F(iF, jF, kF + 1, ex)]) -
-                            TWT *  fh[idx_fh_F(iF, jF, kF,     ex)]
-                        ) / dZ;
-
-                        f_rhs[p] += (eps / cof) * (Dx + Dy + Dz); /* sign=+ */
-                    }
-                }
+                    // Fortran:
+                    // f_rhs(i,j,k) = f_rhs(i,j,k) + eps/cof*(Dx_term + Dy_term + Dz_term)
+                    f_rhs[p] += (eps / cof) * (Dx_term + Dy_term + Dz_term);
             }
         }
-        free(fh);
-        return;
     }
-#elif (ghost_width == 4)
-    /* ---- r=4, cof=256, sign=-, 9pt stencil ---------------------------- */
-    {
-        const int ord = 4;
-        const int r = 4;
-        const double cof = 256.0;
-        const double F8 = 8.0, F28 = 28.0, F56 = 56.0, F70 = 70.0;
-        const int NO_SYMM = 0, EQ_SYMM = 1;
 
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -3;
-
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
-
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
-
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        /* i±4 valid: i-4>=iminF → i0>=iminF+3, i+4<=imaxF → i0<=ex1-5 */
-        const int i0_lo = (iminF + 3 > 0) ? iminF + 3 : 0;
-        const int j0_lo = (jminF + 3 > 0) ? jminF + 3 : 0;
-        const int k0_lo = (kminF + 3 > 0) ? kminF + 3 : 0;
-        const int i0_hi = imaxF - 5;
-        const int j0_hi = jmaxF - 5;
-        const int k0_hi = kmaxF - 5;
-
-        if (!(i0_lo > i0_hi || j0_lo > j0_hi || k0_lo > k0_hi)) {
-            for (int k0 = k0_lo; k0 <= k0_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j0_lo; j0 <= j0_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i0_lo; i0 <= i0_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        /* Stencil: [1,-8,28,-56,70,-56,28,-8,1] */
-                        const double Dx = (
-                            (fh[idx_fh_F_ord4(iF - 4, jF, kF, ex)] + fh[idx_fh_F_ord4(iF + 4, jF, kF, ex)]) -
-                            F8 * (fh[idx_fh_F_ord4(iF - 3, jF, kF, ex)] + fh[idx_fh_F_ord4(iF + 3, jF, kF, ex)]) +
-                            F28* (fh[idx_fh_F_ord4(iF - 2, jF, kF, ex)] + fh[idx_fh_F_ord4(iF + 2, jF, kF, ex)]) -
-                            F56* (fh[idx_fh_F_ord4(iF - 1, jF, kF, ex)] + fh[idx_fh_F_ord4(iF + 1, jF, kF, ex)]) +
-                            F70*  fh[idx_fh_F_ord4(iF,     jF, kF, ex)]
-                        ) / dX;
-
-                        const double Dy = (
-                            (fh[idx_fh_F_ord4(iF, jF - 4, kF, ex)] + fh[idx_fh_F_ord4(iF, jF + 4, kF, ex)]) -
-                            F8 * (fh[idx_fh_F_ord4(iF, jF - 3, kF, ex)] + fh[idx_fh_F_ord4(iF, jF + 3, kF, ex)]) +
-                            F28* (fh[idx_fh_F_ord4(iF, jF - 2, kF, ex)] + fh[idx_fh_F_ord4(iF, jF + 2, kF, ex)]) -
-                            F56* (fh[idx_fh_F_ord4(iF, jF - 1, kF, ex)] + fh[idx_fh_F_ord4(iF, jF + 1, kF, ex)]) +
-                            F70*  fh[idx_fh_F_ord4(iF, jF,     kF, ex)]
-                        ) / dY;
-
-                        const double Dz = (
-                            (fh[idx_fh_F_ord4(iF, jF, kF - 4, ex)] + fh[idx_fh_F_ord4(iF, jF, kF + 4, ex)]) -
-                            F8 * (fh[idx_fh_F_ord4(iF, jF, kF - 3, ex)] + fh[idx_fh_F_ord4(iF, jF, kF + 3, ex)]) +
-                            F28* (fh[idx_fh_F_ord4(iF, jF, kF - 2, ex)] + fh[idx_fh_F_ord4(iF, jF, kF + 2, ex)]) -
-                            F56* (fh[idx_fh_F_ord4(iF, jF, kF - 1, ex)] + fh[idx_fh_F_ord4(iF, jF, kF + 1, ex)]) +
-                            F70*  fh[idx_fh_F_ord4(iF, jF, kF,     ex)]
-                        ) / dZ;
-
-                        f_rhs[p] -= (eps / cof) * (Dx + Dy + Dz); /* sign=- */
-                    }
-                }
-            }
-        }
-        free(fh);
-        return;
-    }
-#elif (ghost_width == 5)
-    /* ---- r=5, cof=1024, sign=+, 11pt stencil ------------------------- */
-    {
-        const int ord = 5;
-        const int r = 5;
-        const double cof = 1024.0;
-        const double F10 = 10.0, F45 = 45.0, F120 = 120.0;
-        const double F210 = 210.0, F252 = 252.0;
-        const int NO_SYMM = 0, EQ_SYMM = 1;
-
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -4;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -4;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -4;
-
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
-
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
-
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        /* i±5 valid: i0>=iminF+4, i0<=ex1-6 */
-        const int i0_lo = (iminF + 4 > 0) ? iminF + 4 : 0;
-        const int j0_lo = (jminF + 4 > 0) ? jminF + 4 : 0;
-        const int k0_lo = (kminF + 4 > 0) ? kminF + 4 : 0;
-        const int i0_hi = imaxF - 6;
-        const int j0_hi = jmaxF - 6;
-        const int k0_hi = kmaxF - 6;
-
-        if (!(i0_lo > i0_hi || j0_lo > j0_hi || k0_lo > k0_hi)) {
-            for (int k0 = k0_lo; k0 <= k0_hi; ++k0) {
-                const int kF = k0 + 1;
-                for (int j0 = j0_lo; j0 <= j0_hi; ++j0) {
-                    const int jF = j0 + 1;
-                    for (int i0 = i0_lo; i0 <= i0_hi; ++i0) {
-                        const int iF = i0 + 1;
-                        const size_t p = idx_ex(i0, j0, k0, ex);
-
-                        /* Stencil: [1,-10,45,-120,210,-252,210,-120,45,-10,1] */
-                        const double Dx = (
-                            (fh[idx_fh_F_ord5(iF - 5, jF, kF, ex)] + fh[idx_fh_F_ord5(iF + 5, jF, kF, ex)]) -
-                            F10 * (fh[idx_fh_F_ord5(iF - 4, jF, kF, ex)] + fh[idx_fh_F_ord5(iF + 4, jF, kF, ex)]) +
-                            F45 * (fh[idx_fh_F_ord5(iF - 3, jF, kF, ex)] + fh[idx_fh_F_ord5(iF + 3, jF, kF, ex)]) -
-                            F120* (fh[idx_fh_F_ord5(iF - 2, jF, kF, ex)] + fh[idx_fh_F_ord5(iF + 2, jF, kF, ex)]) +
-                            F210* (fh[idx_fh_F_ord5(iF - 1, jF, kF, ex)] + fh[idx_fh_F_ord5(iF + 1, jF, kF, ex)]) -
-                            F252*  fh[idx_fh_F_ord5(iF,     jF, kF, ex)]
-                        ) / dX;
-
-                        const double Dy = (
-                            (fh[idx_fh_F_ord5(iF, jF - 5, kF, ex)] + fh[idx_fh_F_ord5(iF, jF + 5, kF, ex)]) -
-                            F10 * (fh[idx_fh_F_ord5(iF, jF - 4, kF, ex)] + fh[idx_fh_F_ord5(iF, jF + 4, kF, ex)]) +
-                            F45 * (fh[idx_fh_F_ord5(iF, jF - 3, kF, ex)] + fh[idx_fh_F_ord5(iF, jF + 3, kF, ex)]) -
-                            F120* (fh[idx_fh_F_ord5(iF, jF - 2, kF, ex)] + fh[idx_fh_F_ord5(iF, jF + 2, kF, ex)]) +
-                            F210* (fh[idx_fh_F_ord5(iF, jF - 1, kF, ex)] + fh[idx_fh_F_ord5(iF, jF + 1, kF, ex)]) -
-                            F252*  fh[idx_fh_F_ord5(iF, jF,     kF, ex)]
-                        ) / dY;
-
-                        const double Dz = (
-                            (fh[idx_fh_F_ord5(iF, jF, kF - 5, ex)] + fh[idx_fh_F_ord5(iF, jF, kF + 5, ex)]) -
-                            F10 * (fh[idx_fh_F_ord5(iF, jF, kF - 4, ex)] + fh[idx_fh_F_ord5(iF, jF, kF + 4, ex)]) +
-                            F45 * (fh[idx_fh_F_ord5(iF, jF, kF - 3, ex)] + fh[idx_fh_F_ord5(iF, jF, kF + 3, ex)]) -
-                            F120* (fh[idx_fh_F_ord5(iF, jF, kF - 2, ex)] + fh[idx_fh_F_ord5(iF, jF, kF + 2, ex)]) +
-                            F210* (fh[idx_fh_F_ord5(iF, jF, kF - 1, ex)] + fh[idx_fh_F_ord5(iF, jF, kF + 1, ex)]) -
-                            F252*  fh[idx_fh_F_ord5(iF, jF, kF,     ex)]
-                        ) / dZ;
-
-                        f_rhs[p] += (eps / cof) * (Dx + Dy + Dz); /* sign=+ */
-                    }
-                }
-            }
-        }
-        free(fh);
-        return;
-    }
-#else
-#error "kodiss_c.C: unsupported ghost_width (must be 2, 3, 4, or 5)"
-#endif
-}
+    free(fh);
+}

AMSS_NCKU_source/kodiss_sh_c.C

@@ -1,136 +0,0 @@
-#include "macrodef.h"
-#include "share_func.h"
-
-/*
- * kodis_sh — Kreiss-Oliger dissipation on shell patches.
- * Same stencil coefficients as Cartesian kodis. Uses symmetry_stbd.
- */
-extern "C" void kodis_sh_(const int ex[3],
-              const double *X, const double *Y, const double *Z,
-              const double *f, double *f_rhs,
-              const double SoAi[2],
-              int Symmetry, double eps, int sst)
-{
-    (void)sst;
-    const double ZEO=0.0;
-    const int ex1=ex[0], ex2=ex[1], ex3=ex[2];
-    const double dX=X[1]-X[0], dY=Y[1]-Y[0], dZ=Z[1]-Z[0];
-    const int imaxF=ex1, jmaxF=ex2, kmaxF=ex3;
-    const double SoA[2]={SoAi[0],SoAi[1]};
-
-#if (ghost_width == 2)
-    {
-        const int ord=2, r=2;
-        const double cof=16.0, F4=4.0, F6=6.0;
-        const int NO_SYMM=0, OCTANT=2;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-1;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-1;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-1;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        double *fh=(double*)malloc(fh_size*sizeof(double));if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const int i0_lo=(iminF+1>0)?iminF+1:0,j0_lo=(jminF+1>0)?jminF+1:0,k0_lo=2;
-        const int i0_hi=imaxF-3,j0_hi=jmaxF-3,k0_hi=kmaxF-3;
-        if(!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)){
-            for(int k0=k0_lo;k0<=k0_hi;++k0){const int kF=k0+1;
-            for(int j0=j0_lo;j0<=j0_hi;++j0){const int jF=j0+1;
-            for(int i0=i0_lo;i0<=i0_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-                const double Dx=((fh[idx_fh_stbd(iF-2,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+2,jF,kF,ord,ex)])-F4*(fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)])+F6*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dX;
-                const double Dy=((fh[idx_fh_stbd(iF,jF-2,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+2,kF,ord,ex)])-F4*(fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)])+F6*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dY;
-                const double Dz=((fh[idx_fh_stbd(iF,jF,kF-2,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+2,ord,ex)])-F4*(fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)])+F6*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dZ;
-                f_rhs[p]-=(eps/cof)*(Dx+Dy+Dz);
-            }}}
-        }
-        free(fh);return;
-    }
-#elif (ghost_width == 3)
-    {
-        const int ord=3, r=3;
-        const double cof=64.0,SIX=6.0,FIT=15.0,TWT=20.0;
-        const int NO_SYMM=0,OCTANT=2;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-2;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-2;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-2;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        double *fh=(double*)malloc(fh_size*sizeof(double));if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const int i0_lo=(iminF+2>0)?iminF+2:0,j0_lo=(jminF+2>0)?jminF+2:0,k0_lo=3;
-        const int i0_hi=imaxF-4,j0_hi=jmaxF-4,k0_hi=kmaxF-4;
-        if(!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)){
-            for(int k0=k0_lo;k0<=k0_hi;++k0){const int kF=k0+1;
-            for(int j0=j0_lo;j0<=j0_hi;++j0){const int jF=j0+1;
-            for(int i0=i0_lo;i0<=i0_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-                const double Dx=((fh[idx_fh_stbd(iF-3,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+3,jF,kF,ord,ex)])-SIX*(fh[idx_fh_stbd(iF-2,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+2,jF,kF,ord,ex)])+FIT*(fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)])-TWT*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dX;
-                const double Dy=((fh[idx_fh_stbd(iF,jF-3,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+3,kF,ord,ex)])-SIX*(fh[idx_fh_stbd(iF,jF-2,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+2,kF,ord,ex)])+FIT*(fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)])-TWT*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dY;
-                const double Dz=((fh[idx_fh_stbd(iF,jF,kF-3,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+3,ord,ex)])-SIX*(fh[idx_fh_stbd(iF,jF,kF-2,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+2,ord,ex)])+FIT*(fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)])-TWT*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dZ;
-                f_rhs[p]+=(eps/cof)*(Dx+Dy+Dz);
-            }}}
-        }
-        free(fh);return;
-    }
-#elif (ghost_width == 4)
-    {
-        const int ord=4, r=4;
-        const double cof=256.0,F8=8.0,F28=28.0,F56=56.0,F70=70.0;
-        const int NO_SYMM=0,OCTANT=2;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-3;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-3;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-3;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        double *fh=(double*)malloc(fh_size*sizeof(double));if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const int i0_lo=(iminF+3>0)?iminF+3:0,j0_lo=(jminF+3>0)?jminF+3:0,k0_lo=4;
-        const int i0_hi=imaxF-5,j0_hi=jmaxF-5,k0_hi=kmaxF-5;
-        if(!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)){
-            for(int k0=k0_lo;k0<=k0_hi;++k0){const int kF=k0+1;
-            for(int j0=j0_lo;j0<=j0_hi;++j0){const int jF=j0+1;
-            for(int i0=i0_lo;i0<=i0_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-                const double Dx=((fh[idx_fh_stbd(iF-4,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+4,jF,kF,ord,ex)])-F8*(fh[idx_fh_stbd(iF-3,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+3,jF,kF,ord,ex)])+F28*(fh[idx_fh_stbd(iF-2,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+2,jF,kF,ord,ex)])-F56*(fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)])+F70*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dX;
-                const double Dy=((fh[idx_fh_stbd(iF,jF-4,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+4,kF,ord,ex)])-F8*(fh[idx_fh_stbd(iF,jF-3,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+3,kF,ord,ex)])+F28*(fh[idx_fh_stbd(iF,jF-2,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+2,kF,ord,ex)])-F56*(fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)])+F70*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dY;
-                const double Dz=((fh[idx_fh_stbd(iF,jF,kF-4,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+4,ord,ex)])-F8*(fh[idx_fh_stbd(iF,jF,kF-3,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+3,ord,ex)])+F28*(fh[idx_fh_stbd(iF,jF,kF-2,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+2,ord,ex)])-F56*(fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)])+F70*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dZ;
-                f_rhs[p]-=(eps/cof)*(Dx+Dy+Dz);
-            }}}
-        }
-        free(fh);return;
-    }
-#elif (ghost_width == 5)
-    {
-        const int ord=5, r=5;
-        const double cof=1024.0,F10=10.0,F45k=45.0,F120=120.0,F210=210.0,F252=252.0;
-        const int NO_SYMM=0,OCTANT=2;
-        int iminF=1,jminF=1,kminF=1;
-        if(Symmetry==OCTANT&&fabs(X[0])<dX)iminF=-4;
-        if(Symmetry==OCTANT&&fabs(Y[0])<dY)jminF=-4;
-        if((sst==2||sst==4)&&fabs(Y[0])<dY)jminF=-4;
-
-        const size_t nx=(size_t)ex1+2*ord,ny=(size_t)ex2+2*ord,nz=(size_t)ex3,fh_size=nx*ny*nz;
-        double *fh=(double*)malloc(fh_size*sizeof(double));if(!fh)return;
-        symmetry_stbd(ord,ex,f,fh,SoA);
-
-        const int i0_lo=(iminF+4>0)?iminF+4:0,j0_lo=(jminF+4>0)?jminF+4:0,k0_lo=5;
-        const int i0_hi=imaxF-6,j0_hi=jmaxF-6,k0_hi=kmaxF-6;
-        if(!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)){
-            for(int k0=k0_lo;k0<=k0_hi;++k0){const int kF=k0+1;
-            for(int j0=j0_lo;j0<=j0_hi;++j0){const int jF=j0+1;
-            for(int i0=i0_lo;i0<=i0_hi;++i0){const int iF=i0+1;const size_t p=idx_ex(i0,j0,k0,ex);
-                const double Dx=((fh[idx_fh_stbd(iF-5,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+5,jF,kF,ord,ex)])-F10*(fh[idx_fh_stbd(iF-4,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+4,jF,kF,ord,ex)])+F45k*(fh[idx_fh_stbd(iF-3,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+3,jF,kF,ord,ex)])-F120*(fh[idx_fh_stbd(iF-2,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+2,jF,kF,ord,ex)])+F210*(fh[idx_fh_stbd(iF-1,jF,kF,ord,ex)]+fh[idx_fh_stbd(iF+1,jF,kF,ord,ex)])-F252*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dX;
-                const double Dy=((fh[idx_fh_stbd(iF,jF-5,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+5,kF,ord,ex)])-F10*(fh[idx_fh_stbd(iF,jF-4,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+4,kF,ord,ex)])+F45k*(fh[idx_fh_stbd(iF,jF-3,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+3,kF,ord,ex)])-F120*(fh[idx_fh_stbd(iF,jF-2,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+2,kF,ord,ex)])+F210*(fh[idx_fh_stbd(iF,jF-1,kF,ord,ex)]+fh[idx_fh_stbd(iF,jF+1,kF,ord,ex)])-F252*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dY;
-                const double Dz=((fh[idx_fh_stbd(iF,jF,kF-5,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+5,ord,ex)])-F10*(fh[idx_fh_stbd(iF,jF,kF-4,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+4,ord,ex)])+F45k*(fh[idx_fh_stbd(iF,jF,kF-3,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+3,ord,ex)])-F120*(fh[idx_fh_stbd(iF,jF,kF-2,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+2,ord,ex)])+F210*(fh[idx_fh_stbd(iF,jF,kF-1,ord,ex)]+fh[idx_fh_stbd(iF,jF,kF+1,ord,ex)])-F252*fh[idx_fh_stbd(iF,jF,kF,ord,ex)])/dZ;
-                f_rhs[p]+=(eps/cof)*(Dx+Dy+Dz);
-            }}}
-        }
-        free(fh);return;
-    }
-#else
-#error "kodiss_sh_c.C: unsupported ghost_width"
-#endif
-}

AMSS_NCKU_source/lopsided_c.C

@@ -1,13 +1,14 @@
-#include "macrodef.h"
 #include "tool.h"
-
 /*
- * C 版 lopsided — upwind (lopsided) advection derivatives.
+ * 你需要提供 symmetry_bd 的 C 版本（或 Fortran 绑到 C 的接口）。
+ * Fortran: call symmetry_bd(3,ex,f,fh,SoA)
  *
- * Adds advection terms to f_rhs for all three spatial directions.
- * Uses sign-biased (one-sided) stencils with centered fallbacks.
- *
- * For lopsided, symmetry_bd ord = ghost_width (same as kodiss).
+ * 约定：
+ *   nghost = 3
+ *   ex[3]  = {ex1,ex2,ex3}
+ *   f      = 原始网格 (ex1*ex2*ex3)
+ *   fh     = 扩展网格 ((ex1+3)*(ex2+3)*(ex3+3))，对应 Fortran 的 (-2:ex1, ...)
+ *   SoA[3] = 输入参数
  */
 void lopsided(const int ex[3],
               const double *X, const double *Y, const double *Z,
@@ -15,577 +16,240 @@ void lopsided(const int ex[3],
               const double *Sfx, const double *Sfy, const double *Sfz,
               int Symmetry, const double SoA[3])
 {
-    const double ZEO = 0.0, ONE = 1.0;
-    const double TWO = 2.0, F6 = 6.0, EIT = 8.0;
-    const double F3 = 3.0, F4 = 4.0, F5 = 5.0, F10 = 10.0, F12 = 12.0, F18 = 18.0;
-    const double F9 = 9.0, F45 = 45.0, F60 = 60.0;
-    const double F2 = 2.0, F15 = 15.0, F24 = 24.0, F30 = 30.0, F35 = 35.0;
-    const double F50 = 50.0, F77 = 77.0, F80 = 80.0, F100 = 100.0, F150 = 150.0;
-    const double F32 = 32.0, F168 = 168.0, F672 = 672.0, F840 = 840.0;
-    const double F140=140.0, F378=378.0, F420=420.0, F1050=1050.0;
+    const double ZEO = 0.0, ONE = 1.0, F3 = 3.0;
+    const double TWO = 2.0, F6 = 6.0, F18 = 18.0;
+    const double F12 = 12.0, F10 = 10.0, EIT = 8.0;
+
+    const int NO_SYMM = 0, EQ_SYMM = 1, OCTANT = 2;
+    (void)OCTANT; // 这里和 Fortran 一样只是定义了不用也没关系
 
-    const int NO_SYMM = 0, EQ_SYMM = 1;
     const int ex1 = ex[0], ex2 = ex[1], ex3 = ex[2];
 
+    // 对应 Fortran: dX = X(2)-X(1)  （Fortran 1-based）
+    // C: X[1]-X[0]
     const double dX = X[1] - X[0];
     const double dY = Y[1] - Y[0];
     const double dZ = Z[1] - Z[0];
 
-#if (ghost_width == 2)
-    /* ---- 2nd-order lopsided --------------------------------------------- */
-    {
-        const int ord = 2;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -1;
+    const double d12dx = ONE / F12 / dX;
+    const double d12dy = ONE / F12 / dY;
+    const double d12dz = ONE / F12 / dZ;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
+    // Fortran 里算了 d2dx/d2dy/d2dz 但本 subroutine 里没用到（保持一致也算出来）
+    const double d2dx  = ONE / TWO / dX;
+    const double d2dy  = ONE / TWO / dY;
+    const double d2dz  = ONE / TWO / dZ;
+    (void)d2dx; (void)d2dy; (void)d2dz;
 
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
+    // Fortran:
+    // imax = ex(1); jmax = ex(2); kmax = ex(3)
+    const int imaxF = ex1;
+    const int jmaxF = ex2;
+    const int kmaxF = ex3;
 
-        const double d2dx  = ONE / TWO / dX;
-        const double d2dy  = ONE / TWO / dY;
-        const double d2dz  = ONE / TWO / dZ;
+    // Fortran:
+    // imin=jmin=kmin=1; 若满足对称条件则设为 -2
+    int iminF = 1, jminF = 1, kminF = 1;
+    if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -2;
+    if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -2;
+    if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -2;
 
-        const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
+    // 分配 fh：大小 (ex1+3)*(ex2+3)*(ex3+3)
+    const size_t nx = (size_t)ex1 + 3;
+    const size_t ny = (size_t)ex2 + 3;
+    const size_t nz = (size_t)ex3 + 3;
+    const size_t fh_size = nx * ny * nz;
 
-        for (int k0 = 0; k0 <= ex3 - 2; ++k0) {
-            const int kF = k0 + 1;
-            for (int j0 = 0; j0 <= ex2 - 2; ++j0) {
-                const int jF = j0 + 1;
-                for (int i0 = 0; i0 <= ex1 - 2; ++i0) {
-                    const int iF = i0 + 1;
-                    const size_t p = idx_ex(i0, j0, k0, ex);
+    double *fh = (double*)malloc(fh_size * sizeof(double));
+    if (!fh) return; // 内存不足：直接返回（你也可以改成 abort/报错）
 
-                    /* x-direction */
-                    const double sfx = Sfx[p];
-                    if (sfx > ZEO) {
-                        if (i0 <= ex1 - 3) // i+2 <= imax
-                            f_rhs[p] += sfx * d2dx * (
-                                -F3*fh[idx_fh_F_ord2(iF,   jF, kF, ex)] +
-                                 F4*fh[idx_fh_F_ord2(iF+1, jF, kF, ex)] -
-                                    fh[idx_fh_F_ord2(iF+2, jF, kF, ex)]);
-                        else if (i0 <= ex1 - 2) // i+1 <= imax
-                            f_rhs[p] += sfx * d2dx * (
-                                -fh[idx_fh_F_ord2(iF,   jF, kF, ex)] +
-                                 fh[idx_fh_F_ord2(iF+1, jF, kF, ex)]);
-                    } else if (sfx < ZEO) {
-                        if ((i0 - 1) >= iminF) // i-2 >= imin
-                            f_rhs[p] -= sfx * d2dx * (
-                                -F3*fh[idx_fh_F_ord2(iF,   jF, kF, ex)] +
-                                 F4*fh[idx_fh_F_ord2(iF-1, jF, kF, ex)] -
-                                    fh[idx_fh_F_ord2(iF-2, jF, kF, ex)]);
-                        else if (i0 >= iminF) // i-1 >= imin
-                            f_rhs[p] -= sfx * d2dx * (
-                                -fh[idx_fh_F_ord2(iF,   jF, kF, ex)] +
-                                 fh[idx_fh_F_ord2(iF-1, jF, kF, ex)]);
+    // Fortran: call symmetry_bd(3,ex,f,fh,SoA)
+    symmetry_bd(3, ex, f, fh, SoA);
+
+    /*
+     * Fortran 主循环：
+     * do k=1,ex(3)-1
+     * do j=1,ex(2)-1
+     * do i=1,ex(1)-1
+     *
+     * 转成 C 0-based：
+     * k0 = 0..ex3-2, j0 = 0..ex2-2, i0 = 0..ex1-2
+     *
+     * 并且 Fortran 里的 i/j/k 在 fh 访问时，仍然是 Fortran 索引值：
+     * iF=i0+1, jF=j0+1, kF=k0+1
+     */
+    for (int k0 = 0; k0 <= ex3 - 2; ++k0) {
+        const int kF = k0 + 1;
+        for (int j0 = 0; j0 <= ex2 - 2; ++j0) {
+            const int jF = j0 + 1;
+            for (int i0 = 0; i0 <= ex1 - 2; ++i0) {
+                const int iF = i0 + 1;
+
+                const size_t p = idx_ex(i0, j0, k0, ex);
+
+                // ---------------- x direction ----------------
+                const double sfx = Sfx[p];
+                if (sfx > ZEO) {
+                    // Fortran: if(i+3 <= imax)
+                    // iF+3 <= ex1  <=> i0+4 <= ex1 <=> i0 <= ex1-4
+                    if (i0 <= ex1 - 4) {
+                        f_rhs[p] += sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF + 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF + 3, jF, kF, ex)]);
                     }
-
-                    /* y-direction */
-                    const double sfy = Sfy[p];
-                    if (sfy > ZEO) {
-                        if (j0 <= ex2-3)
-                            f_rhs[p] += sfy * d2dy * (
-                                -F3*fh[idx_fh_F_ord2(iF, jF,   kF, ex)] +
-                                 F4*fh[idx_fh_F_ord2(iF, jF+1, kF, ex)] -
-                                    fh[idx_fh_F_ord2(iF, jF+2, kF, ex)]);
-                        else if (j0 <= ex2-2)
-                            f_rhs[p] += sfy * d2dy * (
-                                -fh[idx_fh_F_ord2(iF, jF,   kF, ex)] +
-                                 fh[idx_fh_F_ord2(iF, jF+1, kF, ex)]);
-                    } else if (sfy < ZEO) {
-                        if ((j0-1) >= jminF)
-                            f_rhs[p] -= sfy * d2dy * (
-                                -F3*fh[idx_fh_F_ord2(iF, jF,   kF, ex)] +
-                                 F4*fh[idx_fh_F_ord2(iF, jF-1, kF, ex)] -
-                                    fh[idx_fh_F_ord2(iF, jF-2, kF, ex)]);
-                        else if (j0 >= jminF)
-                            f_rhs[p] -= sfy * d2dy * (
-                                -fh[idx_fh_F_ord2(iF, jF,   kF, ex)] +
-                                 fh[idx_fh_F_ord2(iF, jF-1, kF, ex)]);
+                    // elseif(i+2 <= imax)  <=> i0 <= ex1-3
+                    else if (i0 <= ex1 - 3) {
+                        f_rhs[p] += sfx * d12dx *
+                            ( fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -      fh[idx_fh_F(iF + 2, jF, kF, ex)]);
                     }
+                    // elseif(i+1 <= imax)  <=> i0 <= ex1-2（循环里总成立）
+                    else if (i0 <= ex1 - 2) {
+                        f_rhs[p] -= sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF - 3, jF, kF, ex)]);
+                    }
+                } else if (sfx < ZEO) {
+                    // Fortran: if(i-3 >= imin)
+                    // (iF-3) >= iminF  <=> (i0-2) >= iminF
+                    if ((i0 - 2) >= iminF) {
+                        f_rhs[p] -= sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF - 3, jF, kF, ex)]);
+                    }
+                    // elseif(i-2 >= imin) <=> (i0-1) >= iminF
+                    else if ((i0 - 1) >= iminF) {
+                        f_rhs[p] += sfx * d12dx *
+                            ( fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -      fh[idx_fh_F(iF + 2, jF, kF, ex)]);
+                    }
+                    // elseif(i-1 >= imin) <=> i0 >= iminF
+                    else if (i0 >= iminF) {
+                        f_rhs[p] += sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF + 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF + 3, jF, kF, ex)]);
+                    }
+                }
 
-                    /* z-direction */
-                    const double sfz = Sfz[p];
-                    if (sfz > ZEO) {
-                        if (k0 <= ex3-3)
-                            f_rhs[p] += sfz * d2dz * (
-                                -F3*fh[idx_fh_F_ord2(iF, jF, kF,   ex)] +
-                                 F4*fh[idx_fh_F_ord2(iF, jF, kF+1, ex)] -
-                                    fh[idx_fh_F_ord2(iF, jF, kF+2, ex)]);
-                        else if (k0 <= ex3-2)
-                            f_rhs[p] += sfz * d2dz * (
-                                -fh[idx_fh_F_ord2(iF, jF, kF,   ex)] +
-                                 fh[idx_fh_F_ord2(iF, jF, kF+1, ex)]);
-                    } else if (sfz < ZEO) {
-                        if ((k0-1) >= kminF)
-                            f_rhs[p] -= sfz * d2dz * (
-                                -F3*fh[idx_fh_F_ord2(iF, jF, kF,   ex)] +
-                                 F4*fh[idx_fh_F_ord2(iF, jF, kF-1, ex)] -
-                                    fh[idx_fh_F_ord2(iF, jF, kF-2, ex)]);
-                        else if (k0 >= kminF)
-                            f_rhs[p] -= sfz * d2dz * (
-                                -fh[idx_fh_F_ord2(iF, jF, kF,   ex)] +
-                                 fh[idx_fh_F_ord2(iF, jF, kF-1, ex)]);
+                // ---------------- y direction ----------------
+                const double sfy = Sfy[p];
+                if (sfy > ZEO) {
+                    // jF+3 <= ex2 <=> j0+4 <= ex2 <=> j0 <= ex2-4
+                    if (j0 <= ex2 - 4) {
+                        f_rhs[p] += sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF + 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF + 3, kF, ex)]);
+                    } else if (j0 <= ex2 - 3) {
+                        f_rhs[p] += sfy * d12dy *
+                            ( fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -      fh[idx_fh_F(iF, jF + 2, kF, ex)]);
+                    } else if (j0 <= ex2 - 2) {
+                        f_rhs[p] -= sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF - 3, kF, ex)]);
+                    }
+                } else if (sfy < ZEO) {
+                    if ((j0 - 2) >= jminF) {
+                        f_rhs[p] -= sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF - 3, kF, ex)]);
+                    } else if ((j0 - 1) >= jminF) {
+                        f_rhs[p] += sfy * d12dy *
+                            ( fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -      fh[idx_fh_F(iF, jF + 2, kF, ex)]);
+                    } else if (j0 >= jminF) {
+                        f_rhs[p] += sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF + 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF + 3, kF, ex)]);
+                    }
+                }
+
+                // ---------------- z direction ----------------
+                const double sfz = Sfz[p];
+                if (sfz > ZEO) {
+                    if (k0 <= ex3 - 4) {
+                        f_rhs[p] += sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF + 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF + 3, ex)]);
+                    } else if (k0 <= ex3 - 3) {
+                        f_rhs[p] += sfz * d12dz *
+                            ( fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -      fh[idx_fh_F(iF, jF, kF + 2, ex)]);
+                    } else if (k0 <= ex3 - 2) {
+                        f_rhs[p] -= sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF - 3, ex)]);
+                    }
+                } else if (sfz < ZEO) {
+                    if ((k0 - 2) >= kminF) {
+                        f_rhs[p] -= sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF - 3, ex)]);
+                    } else if ((k0 - 1) >= kminF) {
+                        f_rhs[p] += sfz * d12dz *
+                            ( fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -      fh[idx_fh_F(iF, jF, kF + 2, ex)]);
+                    } else if (k0 >= kminF) {
+                        f_rhs[p] += sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF + 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF + 3, ex)]);
                     }
                 }
             }
         }
-        free(fh);
-        return;
     }
-#elif (ghost_width == 3)
-    /* ---- 4th-order lopsided (original code) ---------------------------- */
-    {
-        const int ord = 3;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -2;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -2;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -2;
-
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
-
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d12dx = ONE / F12 / dX;
-        const double d12dy = ONE / F12 / dY;
-        const double d12dz = ONE / F12 / dZ;
-
-        const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
-
-        for (int k0 = 0; k0 <= ex3 - 2; ++k0) {
-            const int kF = k0 + 1;
-            for (int j0 = 0; j0 <= ex2 - 2; ++j0) {
-                const int jF = j0 + 1;
-                for (int i0 = 0; i0 <= ex1 - 2; ++i0) {
-                    const int iF = i0 + 1;
-                    const size_t p = idx_ex(i0, j0, k0, ex);
-
-                    const double sfx = Sfx[p];
-                    if (sfx > ZEO) {
-                        if (i0 <= ex1 - 4) // i+3 <= imax
-                            f_rhs[p] += sfx * d12dx * (
-                                -F3 *fh[idx_fh_F(iF-1, jF, kF, ex)]
-                                -F10*fh[idx_fh_F(iF,   jF, kF, ex)]
-                                +F18*fh[idx_fh_F(iF+1, jF, kF, ex)]
-                                -F6 *fh[idx_fh_F(iF+2, jF, kF, ex)]
-                                +    fh[idx_fh_F(iF+3, jF, kF, ex)]);
-                        else if (i0 <= ex1 - 3) // i+2 <= imax
-                            f_rhs[p] += sfx * d12dx * (
-                                fh[idx_fh_F(iF-2, jF, kF, ex)]
-                                -EIT*fh[idx_fh_F(iF-1, jF, kF, ex)]
-                                +EIT*fh[idx_fh_F(iF+1, jF, kF, ex)]
-                                -    fh[idx_fh_F(iF+2, jF, kF, ex)]);
-                        else if (i0 <= ex1 - 2) // i+1 <= imax → mirrored
-                            f_rhs[p] -= sfx * d12dx * (
-                                -F3 *fh[idx_fh_F(iF+1, jF, kF, ex)]
-                                -F10*fh[idx_fh_F(iF,   jF, kF, ex)]
-                                +F18*fh[idx_fh_F(iF-1, jF, kF, ex)]
-                                -F6 *fh[idx_fh_F(iF-2, jF, kF, ex)]
-                                +    fh[idx_fh_F(iF-3, jF, kF, ex)]);
-                    } else if (sfx < ZEO) {
-                        if ((i0 - 2) >= iminF) // i-3 >= imin
-                            f_rhs[p] -= sfx * d12dx * (
-                                -F3 *fh[idx_fh_F(iF+1, jF, kF, ex)]
-                                -F10*fh[idx_fh_F(iF,   jF, kF, ex)]
-                                +F18*fh[idx_fh_F(iF-1, jF, kF, ex)]
-                                -F6 *fh[idx_fh_F(iF-2, jF, kF, ex)]
-                                +    fh[idx_fh_F(iF-3, jF, kF, ex)]);
-                        else if ((i0 - 1) >= iminF) // i-2 >= imin
-                            f_rhs[p] += sfx * d12dx * (
-                                fh[idx_fh_F(iF-2, jF, kF, ex)]
-                                -EIT*fh[idx_fh_F(iF-1, jF, kF, ex)]
-                                +EIT*fh[idx_fh_F(iF+1, jF, kF, ex)]
-                                -    fh[idx_fh_F(iF+2, jF, kF, ex)]);
-                        else if (i0 >= iminF) // i-1 >= imin → mirrored
-                            f_rhs[p] += sfx * d12dx * (
-                                -F3 *fh[idx_fh_F(iF-1, jF, kF, ex)]
-                                -F10*fh[idx_fh_F(iF,   jF, kF, ex)]
-                                +F18*fh[idx_fh_F(iF+1, jF, kF, ex)]
-                                -F6 *fh[idx_fh_F(iF+2, jF, kF, ex)]
-                                +    fh[idx_fh_F(iF+3, jF, kF, ex)]);
-                    }
-
-                    const double sfy = Sfy[p];
-                    if (sfy > ZEO) {
-                        if (j0 <= ex2-4)
-                            f_rhs[p] += sfy * d12dy * (
-                                -F3*fh[idx_fh_F(iF,jF-1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF+1,kF,ex)]-F6*fh[idx_fh_F(iF,jF+2,kF,ex)]
-                                +fh[idx_fh_F(iF,jF+3,kF,ex)]);
-                        else if (j0 <= ex2-3)
-                            f_rhs[p] += sfy * d12dy * (fh[idx_fh_F(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F(iF,jF+1,kF,ex)]-fh[idx_fh_F(iF,jF+2,kF,ex)]);
-                        else if (j0 <= ex2-2)
-                            f_rhs[p] -= sfy * d12dy * (
-                                -F3*fh[idx_fh_F(iF,jF+1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF-1,kF,ex)]-F6*fh[idx_fh_F(iF,jF-2,kF,ex)]
-                                +fh[idx_fh_F(iF,jF-3,kF,ex)]);
-                    } else if (sfy < ZEO) {
-                        if ((j0-2) >= jminF)
-                            f_rhs[p] -= sfy * d12dy * (
-                                -F3*fh[idx_fh_F(iF,jF+1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF-1,kF,ex)]-F6*fh[idx_fh_F(iF,jF-2,kF,ex)]
-                                +fh[idx_fh_F(iF,jF-3,kF,ex)]);
-                        else if ((j0-1) >= jminF)
-                            f_rhs[p] += sfy * d12dy * (fh[idx_fh_F(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F(iF,jF+1,kF,ex)]-fh[idx_fh_F(iF,jF+2,kF,ex)]);
-                        else if (j0 >= jminF)
-                            f_rhs[p] += sfy * d12dy * (
-                                -F3*fh[idx_fh_F(iF,jF-1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF+1,kF,ex)]-F6*fh[idx_fh_F(iF,jF+2,kF,ex)]
-                                +fh[idx_fh_F(iF,jF+3,kF,ex)]);
-                    }
-
-                    const double sfz = Sfz[p];
-                    if (sfz > ZEO) {
-                        if (k0 <= ex3-4)
-                            f_rhs[p] += sfz * d12dz * (
-                                -F3*fh[idx_fh_F(iF,jF,kF-1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF,kF+1,ex)]-F6*fh[idx_fh_F(iF,jF,kF+2,ex)]
-                                +fh[idx_fh_F(iF,jF,kF+3,ex)]);
-                        else if (k0 <= ex3-3)
-                            f_rhs[p] += sfz * d12dz * (fh[idx_fh_F(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F(iF,jF,kF+1,ex)]-fh[idx_fh_F(iF,jF,kF+2,ex)]);
-                        else if (k0 <= ex3-2)
-                            f_rhs[p] -= sfz * d12dz * (
-                                -F3*fh[idx_fh_F(iF,jF,kF+1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF,kF-1,ex)]-F6*fh[idx_fh_F(iF,jF,kF-2,ex)]
-                                +fh[idx_fh_F(iF,jF,kF-3,ex)]);
-                    } else if (sfz < ZEO) {
-                        if ((k0-2) >= kminF)
-                            f_rhs[p] -= sfz * d12dz * (
-                                -F3*fh[idx_fh_F(iF,jF,kF+1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF,kF-1,ex)]-F6*fh[idx_fh_F(iF,jF,kF-2,ex)]
-                                +fh[idx_fh_F(iF,jF,kF-3,ex)]);
-                        else if ((k0-1) >= kminF)
-                            f_rhs[p] += sfz * d12dz * (fh[idx_fh_F(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F(iF,jF,kF+1,ex)]-fh[idx_fh_F(iF,jF,kF+2,ex)]);
-                        else if (k0 >= kminF)
-                            f_rhs[p] += sfz * d12dz * (
-                                -F3*fh[idx_fh_F(iF,jF,kF-1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]
-                                +F18*fh[idx_fh_F(iF,jF,kF+1,ex)]-F6*fh[idx_fh_F(iF,jF,kF+2,ex)]
-                                +fh[idx_fh_F(iF,jF,kF+3,ex)]);
-                    }
-                }
-            }
-        }
-        free(fh);
-        return;
-    }
-#elif (ghost_width == 4)
-    /* ---- 6th-order lopsided --------------------------------------------- */
-    {
-        const int ord = 4;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -3;
-
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
-
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d60dx = ONE / F60 / dX;
-        const double d60dy = ONE / F60 / dY;
-        const double d60dz = ONE / F60 / dZ;
-        const double d12dx = ONE / F12 / dX;
-        const double d12dy = ONE / F12 / dY;
-        const double d12dz = ONE / F12 / dZ;
-        const double d2dx  = ONE / TWO / dX;
-        const double d2dy  = ONE / TWO / dY;
-        const double d2dz  = ONE / TWO / dZ;
-
-        const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
-
-        for (int k0 = 0; k0 <= ex3 - 2; ++k0) {
-            const int kF = k0 + 1;
-            for (int j0 = 0; j0 <= ex2 - 2; ++j0) {
-                const int jF = j0 + 1;
-                for (int i0 = 0; i0 <= ex1 - 2; ++i0) {
-                    const int iF = i0 + 1;
-                    const size_t p = idx_ex(i0, j0, k0, ex);
-
-                    /* ---- x-direction ---- */
-                    const double sfx = Sfx[p];
-                    if (sfx > ZEO) {
-                        /* Primary biased: 2*f(i-2)-24*f(i-1)-35*f(i)+80*f(i+1)-30*f(i+2)+8*f(i+3)-f(i+4) */
-                        if (i0 <= ex1-5 && (i0-1)>=iminF) // i+4<=imax && i-2>=imin
-                            f_rhs[p] += sfx * d60dx * (
-                                +F2*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-F24*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]
-                                -F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]
-                                -F30*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]
-                                -fh[idx_fh_F_ord4(iF+4,jF,kF,ex)]);
-                        /* Boundary-adapted: -10*f(i-1)-77*f(i)+150*f(i+1)-100*f(i+2)+50*f(i+3)-15*f(i+4)+2*f(i+5) */
-                        else if (i0 <= ex1-6 && i0 >= iminF) // i+5<=imax && i-1>=imin
-                            f_rhs[p] += sfx * d60dx * (
-                                -F10*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]
-                                +F150*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F100*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]
-                                +F50*fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]-F15*fh[idx_fh_F_ord4(iF+4,jF,kF,ex)]
-                                +F2*fh[idx_fh_F_ord4(iF+5,jF,kF,ex)]);
-                        /* Centered fallbacks */
-                        else if (i0 <= ex1-4 && (i0-2)>=iminF) // 6th: i+3<=imax && i-3>=imin
-                            f_rhs[p] += sfx * d60dx * (
-                                -fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]
-                                -F45*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]
-                                -F9*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]);
-                        else if (i0 <= ex1-3 && (i0-1)>=iminF) // 4th
-                            f_rhs[p] += sfx * d12dx * (
-                                fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]
-                                +EIT*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]);
-                        else if (i0 <= ex1-2 && i0>=iminF) // 2nd
-                            f_rhs[p] += sfx * d2dx * (
-                                -fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]);
-                    } else if (sfx < ZEO) {
-                        if ((i0-4)>=iminF && i0<=ex1-2) // i-4>=imin && i+2<=imax
-                            f_rhs[p] -= sfx * d60dx * (
-                                +F2*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]-F24*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]
-                                -F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]
-                                -F30*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]
-                                -fh[idx_fh_F_ord4(iF-4,jF,kF,ex)]);
-                        else if ((i0-5)>=iminF && i0<=ex1-2) // i-5>=imin && i+1<=imax
-                            f_rhs[p] -= sfx * d60dx * (
-                                -F10*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]
-                                +F150*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]-F100*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]
-                                +F50*fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]-F15*fh[idx_fh_F_ord4(iF-4,jF,kF,ex)]
-                                +F2*fh[idx_fh_F_ord4(iF-5,jF,kF,ex)]);
-                        else if ((i0-3)>=iminF && i0<=ex1-2) // 6th centered
-                            f_rhs[p] -= sfx * d60dx * (
-                                -fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]
-                                -F45*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]
-                                -F9*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]);
-                        else if ((i0-2)>=iminF && i0<=ex1-2) // 4th
-                            f_rhs[p] -= sfx * d12dx * (
-                                fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]
-                                +EIT*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]);
-                        else if ((i0-1)>=iminF && i0<=ex1-2) // 2nd
-                            f_rhs[p] -= sfx * d2dx * (
-                                -fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]);
-                    }
-
-                    /* ---- y-direction ---- */
-                    const double sfy = Sfy[p];
-                    if (sfy > ZEO) {
-                        if (j0<=ex2-5 && (j0-1)>=jminF)
-                            f_rhs[p] += sfy * d60dy*(F2*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-F24*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F30*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]-fh[idx_fh_F_ord4(iF,jF+4,kF,ex)]);
-                        else if (j0<=ex2-6 && j0>=jminF)
-                            f_rhs[p] += sfy * d60dy*(-F10*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F100*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+F50*fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]-F15*fh[idx_fh_F_ord4(iF,jF+4,kF,ex)]+F2*fh[idx_fh_F_ord4(iF,jF+5,kF,ex)]);
-                        else if (j0<=ex2-4 && (j0-2)>=jminF)
-                            f_rhs[p] += sfy * d60dy*(-fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]);
-                        else if (j0<=ex2-3 && (j0-1)>=jminF)
-                            f_rhs[p] += sfy * d12dy*(fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]);
-                        else if (j0<=ex2-2 && j0>=jminF)
-                            f_rhs[p] += sfy * d2dy*(-fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]);
-                    } else if (sfy < ZEO) {
-                        if ((j0-4)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d60dy*(F2*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]-F24*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F30*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]-fh[idx_fh_F_ord4(iF,jF-4,kF,ex)]);
-                        else if ((j0-5)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d60dy*(-F10*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F100*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]+F50*fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]-F15*fh[idx_fh_F_ord4(iF,jF-4,kF,ex)]+F2*fh[idx_fh_F_ord4(iF,jF-5,kF,ex)]);
-                        else if ((j0-3)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d60dy*(-fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]);
-                        else if ((j0-2)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d12dy*(fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]);
-                        else if ((j0-1)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d2dy*(-fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]);
-                    }
-
-                    /* ---- z-direction ---- */
-                    const double sfz = Sfz[p];
-                    if (sfz > ZEO) {
-                        if (k0<=ex3-5 && (k0-1)>=kminF)
-                            f_rhs[p] += sfz * d60dz*(F2*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-F24*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F30*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]-fh[idx_fh_F_ord4(iF,jF,kF+4,ex)]);
-                        else if (k0<=ex3-6 && k0>=kminF)
-                            f_rhs[p] += sfz * d60dz*(-F10*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F100*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+F50*fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]-F15*fh[idx_fh_F_ord4(iF,jF,kF+4,ex)]+F2*fh[idx_fh_F_ord4(iF,jF,kF+5,ex)]);
-                        else if (k0<=ex3-4 && (k0-2)>=kminF)
-                            f_rhs[p] += sfz * d60dz*(-fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]);
-                        else if (k0<=ex3-3 && (k0-1)>=kminF)
-                            f_rhs[p] += sfz * d12dz*(fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]);
-                        else if (k0<=ex3-2 && k0>=kminF)
-                            f_rhs[p] += sfz * d2dz*(-fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]);
-                    } else if (sfz < ZEO) {
-                        if ((k0-4)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d60dz*(F2*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]-F24*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F30*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]-fh[idx_fh_F_ord4(iF,jF,kF-4,ex)]);
-                        else if ((k0-5)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d60dz*(-F10*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F100*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]+F50*fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]-F15*fh[idx_fh_F_ord4(iF,jF,kF-4,ex)]+F2*fh[idx_fh_F_ord4(iF,jF,kF-5,ex)]);
-                        else if ((k0-3)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d60dz*(-fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]);
-                        else if ((k0-2)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d12dz*(fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]);
-                        else if ((k0-1)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d2dz*(-fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]);
-                    }
-                }
-            }
-        }
-        free(fh);
-        return;
-    }
-#elif (ghost_width == 5)
-    /* ---- 8th-order lopsided --------------------------------------------- */
-    {
-        const int ord = 5;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -4;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -4;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -4;
-
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        const size_t fh_size = nx * ny * nz;
-
-        double *fh = (double*)malloc(fh_size * sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d840dx = ONE / F840 / dX;
-        const double d840dy = ONE / F840 / dY;
-        const double d840dz = ONE / F840 / dZ;
-        const double d60dx = ONE / F60 / dX;
-        const double d60dy = ONE / F60 / dY;
-        const double d60dz = ONE / F60 / dZ;
-        const double d12dx = ONE / F12 / dX;
-        const double d12dy = ONE / F12 / dY;
-        const double d12dz = ONE / F12 / dZ;
-        const double d2dx  = ONE / TWO / dX;
-        const double d2dy  = ONE / TWO / dY;
-        const double d2dz  = ONE / TWO / dZ;
-
-        const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
-
-        for (int k0 = 0; k0 <= ex3 - 2; ++k0) {
-            const int kF = k0 + 1;
-            for (int j0 = 0; j0 <= ex2 - 2; ++j0) {
-                const int jF = j0 + 1;
-                for (int i0 = 0; i0 <= ex1 - 2; ++i0) {
-                    const int iF = i0 + 1;
-                    const size_t p = idx_ex(i0, j0, k0, ex);
-
-                    const double sfx = Sfx[p];
-                    if (sfx > ZEO) {
-                        /* 8th biased: -5*f(i-3)+60*f(i-2)-420*f(i-1)-378*f(i)+1050*f(i+1)-420*f(i+2)+140*f(i+3)-30*f(i+4)+3*f(i+5) */
-                        if (i0 <= ex1-6 && (i0-2)>=iminF) // i+5<=imax && i-3>=imin
-                            f_rhs[p] += sfx * d840dx * (
-                                -F5*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F60*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]
-                                -F420*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]
-                                +F1050*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F420*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]
-                                +F140*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]-F30*fh[idx_fh_F_ord5(iF+4,jF,kF,ex)]
-                                +F3*fh[idx_fh_F_ord5(iF+5,jF,kF,ex)]);
-                        /* 8th centered: +3*f(i-4)-32*f(i-3)+168*f(i-2)-672*f(i-1)+672*f(i+1)-168*f(i+2)+32*f(i+3)-3*f(i+4) */
-                        else if (i0 <= ex1-5 && (i0-3)>=iminF)
-                            f_rhs[p] += sfx * d840dx * (
-                                +F3*fh[idx_fh_F_ord5(iF-4,jF,kF,ex)]-F32*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]
-                                +F168*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-F672*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]
-                                +F672*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F168*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]
-                                +F32*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]-F3*fh[idx_fh_F_ord5(iF+4,jF,kF,ex)]);
-                        else if (i0 <= ex1-4 && (i0-2)>=iminF) // 6th centered
-                            f_rhs[p] += sfx * d60dx * (
-                                -fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]
-                                -F45*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]
-                                -F9*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]);
-                        else if (i0 <= ex1-3 && (i0-1)>=iminF) // 4th centered
-                            f_rhs[p] += sfx * d12dx * (
-                                fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]
-                                +EIT*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]);
-                        else if (i0 <= ex1-2 && i0>=iminF) // 2nd centered
-                            f_rhs[p] += sfx * d2dx * (
-                                -fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]);
-                    } else if (sfx < ZEO) {
-                        if ((i0-5)>=iminF && i0<=ex1-2) // i-5>=imin && i+3<=imax
-                            f_rhs[p] -= sfx * d840dx * (
-                                -F5*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]+F60*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]
-                                -F420*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]
-                                +F1050*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]-F420*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]
-                                +F140*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]-F30*fh[idx_fh_F_ord5(iF-4,jF,kF,ex)]
-                                +F3*fh[idx_fh_F_ord5(iF-5,jF,kF,ex)]);
-                        else if ((i0-4)>=iminF && i0<=ex1-2) // 8th centered
-                            f_rhs[p] -= sfx * d840dx * (
-                                +F3*fh[idx_fh_F_ord5(iF-4,jF,kF,ex)]-F32*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]
-                                +F168*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-F672*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]
-                                +F672*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F168*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]
-                                +F32*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]-F3*fh[idx_fh_F_ord5(iF+4,jF,kF,ex)]);
-                        else if ((i0-3)>=iminF && i0<=ex1-2) // 6th centered
-                            f_rhs[p] -= sfx * d60dx * (
-                                -fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]
-                                -F45*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]
-                                -F9*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]);
-                        else if ((i0-2)>=iminF && i0<=ex1-2) // 4th centered
-                            f_rhs[p] -= sfx * d12dx * (
-                                fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]
-                                +EIT*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]);
-                        else if ((i0-1)>=iminF && i0<=ex1-2) // 2nd centered
-                            f_rhs[p] -= sfx * d2dx * (
-                                -fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]);
-                    }
-
-                    const double sfy = Sfy[p];
-                    if (sfy > ZEO) {
-                        if (j0<=ex2-6 && (j0-2)>=jminF)
-                            f_rhs[p] += sfy * d840dy*(-F5*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F60*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+F140*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]-F30*fh[idx_fh_F_ord5(iF,jF+4,kF,ex)]+F3*fh[idx_fh_F_ord5(iF,jF+5,kF,ex)]);
-                        else if (j0<=ex2-5 && (j0-3)>=jminF)
-                            f_rhs[p] += sfy * d840dy*(+F3*fh[idx_fh_F_ord5(iF,jF-4,kF,ex)]-F32*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F168*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F672*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F672*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F168*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+F32*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]-F3*fh[idx_fh_F_ord5(iF,jF+4,kF,ex)]);
-                        else if (j0<=ex2-4 && (j0-2)>=jminF)
-                            f_rhs[p] += sfy * d60dy*(-fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]);
-                        else if (j0<=ex2-3 && (j0-1)>=jminF)
-                            f_rhs[p] += sfy * d12dy*(fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]);
-                        else if (j0<=ex2-2 && j0>=jminF)
-                            f_rhs[p] += sfy * d2dy*(-fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]);
-                    } else if (sfy < ZEO) {
-                        if ((j0-5)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d840dy*(-F5*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]+F60*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]+F140*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]-F30*fh[idx_fh_F_ord5(iF,jF-4,kF,ex)]+F3*fh[idx_fh_F_ord5(iF,jF-5,kF,ex)]);
-                        else if ((j0-4)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d840dy*(+F3*fh[idx_fh_F_ord5(iF,jF-4,kF,ex)]-F32*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F168*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F672*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F672*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F168*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+F32*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]-F3*fh[idx_fh_F_ord5(iF,jF+4,kF,ex)]);
-                        else if ((j0-3)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d60dy*(-fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]);
-                        else if ((j0-2)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d12dy*(fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]);
-                        else if ((j0-1)>=jminF && j0<=ex2-2)
-                            f_rhs[p] -= sfy * d2dy*(-fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]);
-                    }
-
-                    const double sfz = Sfz[p];
-                    if (sfz > ZEO) {
-                        if (k0<=ex3-6 && (k0-2)>=kminF)
-                            f_rhs[p] += sfz * d840dz*(-F5*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F60*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+F140*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]-F30*fh[idx_fh_F_ord5(iF,jF,kF+4,ex)]+F3*fh[idx_fh_F_ord5(iF,jF,kF+5,ex)]);
-                        else if (k0<=ex3-5 && (k0-3)>=kminF)
-                            f_rhs[p] += sfz * d840dz*(+F3*fh[idx_fh_F_ord5(iF,jF,kF-4,ex)]-F32*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F168*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F672*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F672*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F168*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+F32*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]-F3*fh[idx_fh_F_ord5(iF,jF,kF+4,ex)]);
-                        else if (k0<=ex3-4 && (k0-2)>=kminF)
-                            f_rhs[p] += sfz * d60dz*(-fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]);
-                        else if (k0<=ex3-3 && (k0-1)>=kminF)
-                            f_rhs[p] += sfz * d12dz*(fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]);
-                        else if (k0<=ex3-2 && k0>=kminF)
-                            f_rhs[p] += sfz * d2dz*(-fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]);
-                    } else if (sfz < ZEO) {
-                        if ((k0-5)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d840dz*(-F5*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]+F60*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]+F140*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]-F30*fh[idx_fh_F_ord5(iF,jF,kF-4,ex)]+F3*fh[idx_fh_F_ord5(iF,jF,kF-5,ex)]);
-                        else if ((k0-4)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d840dz*(+F3*fh[idx_fh_F_ord5(iF,jF,kF-4,ex)]-F32*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F168*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F672*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F672*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F168*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+F32*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]-F3*fh[idx_fh_F_ord5(iF,jF,kF+4,ex)]);
-                        else if ((k0-3)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d60dz*(-fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]);
-                        else if ((k0-2)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d12dz*(fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]);
-                        else if ((k0-1)>=kminF && k0<=ex3-2)
-                            f_rhs[p] -= sfz * d2dz*(-fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]);
-                    }
-                }
-            }
-        }
-        free(fh);
-        return;
-    }
-#else
-#error "lopsided_c.C: unsupported ghost_width (must be 2, 3, 4, or 5)"
-#endif
+    free(fh);
 }
+
+
+
+
+

AMSS_NCKU_source/lopsided_kodis_c.C

@@ -1,15 +1,8 @@
-#include "macrodef.h"
 #include "tool.h"
 
 /*
- * C 版 lopsided_kodis — combined upwind advection + KO dissipation.
- * Uses one shared symmetry_bd buffer (ord = ghost_width for both components).
- *
- * FD order selection via ghost_width:
- *   2 → 2nd-order advection + r=2 KO (cof=16, sign=-)
- *   3 → 4th-order advection + r=3 KO (cof=64, sign=+)
- *   4 → 6th-order advection + r=4 KO (cof=256, sign=-)
- *   5 → 8th-order advection + r=5 KO (cof=1024, sign=+)
+ * Combined advection (lopsided) + KO dissipation (kodis).
+ * Uses one shared symmetry_bd buffer per call.
  */
 void lopsided_kodis(const int ex[3],
                     const double *X, const double *Y, const double *Z,
@@ -17,370 +10,239 @@ void lopsided_kodis(const int ex[3],
                     const double *Sfx, const double *Sfy, const double *Sfz,
                     int Symmetry, const double SoA[3], double eps)
 {
-    const double ZEO = 0.0, ONE = 1.0;
-    const double TWO = 2.0, F6 = 6.0, EIT = 8.0;
-    const double F3 = 3.0, F4 = 4.0, F5 = 5.0, F10 = 10.0, F12 = 12.0, F18 = 18.0;
-    const double F9 = 9.0, F45 = 45.0, F60 = 60.0;
-    const double F2 = 2.0, F15 = 15.0, F24 = 24.0, F30 = 30.0, F35 = 35.0;
-    const double F50 = 50.0, F77 = 77.0, F80 = 80.0, F100 = 100.0, F150 = 150.0;
-    const double F32 = 32.0, F168 = 168.0, F672 = 672.0, F840 = 840.0;
-    const double F140=140.0, F378=378.0, F420=420.0, F1050=1050.0;
+    const double ZEO = 0.0, ONE = 1.0, F3 = 3.0;
+    const double F6 = 6.0, F18 = 18.0;
+    const double F12 = 12.0, F10 = 10.0, EIT = 8.0;
+    const double SIX = 6.0, FIT = 15.0, TWT = 20.0;
+    const double cof = 64.0; // 2^6
 
     const int NO_SYMM = 0, EQ_SYMM = 1;
+
     const int ex1 = ex[0], ex2 = ex[1], ex3 = ex[2];
 
     const double dX = X[1] - X[0];
     const double dY = Y[1] - Y[0];
     const double dZ = Z[1] - Z[0];
 
-    const int imaxF = ex1, jmaxF = ex2, kmaxF = ex3;
+    const double d12dx = ONE / F12 / dX;
+    const double d12dy = ONE / F12 / dY;
+    const double d12dz = ONE / F12 / dZ;
 
-#if (ghost_width == 2)
-    {
-        const int ord = 2;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -1;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -1;
+    const int imaxF = ex1;
+    const int jmaxF = ex2;
+    const int kmaxF = ex3;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        double *fh = (double*)malloc(nx*ny*nz*sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
+    int iminF = 1, jminF = 1, kminF = 1;
+    if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -2;
+    if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -2;
+    if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -2;
 
-        const double d2dx = ONE/TWO/dX, d2dy = ONE/TWO/dY, d2dz = ONE/TWO/dZ;
+    // fh for Fortran-style domain (-2:ex1,-2:ex2,-2:ex3)
+    const size_t nx = (size_t)ex1 + 3;
+    const size_t ny = (size_t)ex2 + 3;
+    const size_t nz = (size_t)ex3 + 3;
+    const size_t fh_size = nx * ny * nz;
 
-        /* ---- advection (2nd-order) ---- */
-        for (int k0 = 0; k0 <= ex3-2; ++k0) {
-            const int kF = k0+1;
-            for (int j0 = 0; j0 <= ex2-2; ++j0) {
-                const int jF = j0+1;
-                for (int i0 = 0; i0 <= ex1-2; ++i0) {
-                    const int iF = i0+1;
-                    const size_t p = idx_ex(i0,j0,k0,ex);
+    double *fh = (double*)malloc(fh_size * sizeof(double));
+    if (!fh) return;
 
-                    const double sfx = Sfx[p];
-                    if (sfx > ZEO) {
-                        if (i0<=ex1-3) f_rhs[p] += sfx*d2dx*(-F3*fh[idx_fh_F_ord2(iF,jF,kF,ex)]+F4*fh[idx_fh_F_ord2(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord2(iF+2,jF,kF,ex)]);
-                        else if (i0<=ex1-2) f_rhs[p] += sfx*d2dx*(-fh[idx_fh_F_ord2(iF,jF,kF,ex)]+fh[idx_fh_F_ord2(iF+1,jF,kF,ex)]);
-                    } else if (sfx < ZEO) {
-                        if ((i0-1)>=iminF) f_rhs[p] -= sfx*d2dx*(-F3*fh[idx_fh_F_ord2(iF,jF,kF,ex)]+F4*fh[idx_fh_F_ord2(iF-1,jF,kF,ex)]-fh[idx_fh_F_ord2(iF-2,jF,kF,ex)]);
-                        else if (i0>=iminF) f_rhs[p] -= sfx*d2dx*(-fh[idx_fh_F_ord2(iF,jF,kF,ex)]+fh[idx_fh_F_ord2(iF-1,jF,kF,ex)]);
+    symmetry_bd(3, ex, f, fh, SoA);
+
+    // Advection (same stencil logic as lopsided_c.C)
+    for (int k0 = 0; k0 <= ex3 - 2; ++k0) {
+        const int kF = k0 + 1;
+        for (int j0 = 0; j0 <= ex2 - 2; ++j0) {
+            const int jF = j0 + 1;
+            for (int i0 = 0; i0 <= ex1 - 2; ++i0) {
+                const int iF = i0 + 1;
+                const size_t p = idx_ex(i0, j0, k0, ex);
+
+                const double sfx = Sfx[p];
+                if (sfx > ZEO) {
+                    if (i0 <= ex1 - 4) {
+                        f_rhs[p] += sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF + 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF + 3, jF, kF, ex)]);
+                    } else if (i0 <= ex1 - 3) {
+                        f_rhs[p] += sfx * d12dx *
+                            ( fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -      fh[idx_fh_F(iF + 2, jF, kF, ex)]);
+                    } else if (i0 <= ex1 - 2) {
+                        f_rhs[p] -= sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF - 3, jF, kF, ex)]);
                     }
-                    const double sfy = Sfy[p];
-                    if (sfy > ZEO) {
-                        if (j0<=ex2-3) f_rhs[p] += sfy*d2dy*(-F3*fh[idx_fh_F_ord2(iF,jF,kF,ex)]+F4*fh[idx_fh_F_ord2(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord2(iF,jF+2,kF,ex)]);
-                        else if (j0<=ex2-2) f_rhs[p] += sfy*d2dy*(-fh[idx_fh_F_ord2(iF,jF,kF,ex)]+fh[idx_fh_F_ord2(iF,jF+1,kF,ex)]);
-                    } else if (sfy < ZEO) {
-                        if ((j0-1)>=jminF) f_rhs[p] -= sfy*d2dy*(-F3*fh[idx_fh_F_ord2(iF,jF,kF,ex)]+F4*fh[idx_fh_F_ord2(iF,jF-1,kF,ex)]-fh[idx_fh_F_ord2(iF,jF-2,kF,ex)]);
-                        else if (j0>=jminF) f_rhs[p] -= sfy*d2dy*(-fh[idx_fh_F_ord2(iF,jF,kF,ex)]+fh[idx_fh_F_ord2(iF,jF-1,kF,ex)]);
+                } else if (sfx < ZEO) {
+                    if ((i0 - 2) >= iminF) {
+                        f_rhs[p] -= sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF - 3, jF, kF, ex)]);
+                    } else if ((i0 - 1) >= iminF) {
+                        f_rhs[p] += sfx * d12dx *
+                            ( fh[idx_fh_F(iF - 2, jF, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -      fh[idx_fh_F(iF + 2, jF, kF, ex)]);
+                    } else if (i0 >= iminF) {
+                        f_rhs[p] += sfx * d12dx *
+                            (-F3  * fh[idx_fh_F(iF - 1, jF, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF    , jF, kF, ex)]
+                             +F18 * fh[idx_fh_F(iF + 1, jF, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF + 2, jF, kF, ex)]
+                             +      fh[idx_fh_F(iF + 3, jF, kF, ex)]);
                     }
-                    const double sfz = Sfz[p];
-                    if (sfz > ZEO) {
-                        if (k0<=ex3-3) f_rhs[p] += sfz*d2dz*(-F3*fh[idx_fh_F_ord2(iF,jF,kF,ex)]+F4*fh[idx_fh_F_ord2(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord2(iF,jF,kF+2,ex)]);
-                        else if (k0<=ex3-2) f_rhs[p] += sfz*d2dz*(-fh[idx_fh_F_ord2(iF,jF,kF,ex)]+fh[idx_fh_F_ord2(iF,jF,kF+1,ex)]);
-                    } else if (sfz < ZEO) {
-                        if ((k0-1)>=kminF) f_rhs[p] -= sfz*d2dz*(-F3*fh[idx_fh_F_ord2(iF,jF,kF,ex)]+F4*fh[idx_fh_F_ord2(iF,jF,kF-1,ex)]-fh[idx_fh_F_ord2(iF,jF,kF-2,ex)]);
-                        else if (k0>=kminF) f_rhs[p] -= sfz*d2dz*(-fh[idx_fh_F_ord2(iF,jF,kF,ex)]+fh[idx_fh_F_ord2(iF,jF,kF-1,ex)]);
+                }
+
+                const double sfy = Sfy[p];
+                if (sfy > ZEO) {
+                    if (j0 <= ex2 - 4) {
+                        f_rhs[p] += sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF + 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF + 3, kF, ex)]);
+                    } else if (j0 <= ex2 - 3) {
+                        f_rhs[p] += sfy * d12dy *
+                            ( fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -      fh[idx_fh_F(iF, jF + 2, kF, ex)]);
+                    } else if (j0 <= ex2 - 2) {
+                        f_rhs[p] -= sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF - 3, kF, ex)]);
+                    }
+                } else if (sfy < ZEO) {
+                    if ((j0 - 2) >= jminF) {
+                        f_rhs[p] -= sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF - 3, kF, ex)]);
+                    } else if ((j0 - 1) >= jminF) {
+                        f_rhs[p] += sfy * d12dy *
+                            ( fh[idx_fh_F(iF, jF - 2, kF, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -      fh[idx_fh_F(iF, jF + 2, kF, ex)]);
+                    } else if (j0 >= jminF) {
+                        f_rhs[p] += sfy * d12dy *
+                            (-F3  * fh[idx_fh_F(iF, jF - 1, kF, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF    , kF, ex)]
+                             +F18 * fh[idx_fh_F(iF, jF + 1, kF, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF + 2, kF, ex)]
+                             +      fh[idx_fh_F(iF, jF + 3, kF, ex)]);
+                    }
+                }
+
+                const double sfz = Sfz[p];
+                if (sfz > ZEO) {
+                    if (k0 <= ex3 - 4) {
+                        f_rhs[p] += sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF + 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF + 3, ex)]);
+                    } else if (k0 <= ex3 - 3) {
+                        f_rhs[p] += sfz * d12dz *
+                            ( fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -      fh[idx_fh_F(iF, jF, kF + 2, ex)]);
+                    } else if (k0 <= ex3 - 2) {
+                        f_rhs[p] -= sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF - 3, ex)]);
+                    }
+                } else if (sfz < ZEO) {
+                    if ((k0 - 2) >= kminF) {
+                        f_rhs[p] -= sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF - 3, ex)]);
+                    } else if ((k0 - 1) >= kminF) {
+                        f_rhs[p] += sfz * d12dz *
+                            ( fh[idx_fh_F(iF, jF, kF - 2, ex)]
+                             -EIT * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             +EIT * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -      fh[idx_fh_F(iF, jF, kF + 2, ex)]);
+                    } else if (k0 >= kminF) {
+                        f_rhs[p] += sfz * d12dz *
+                            (-F3  * fh[idx_fh_F(iF, jF, kF - 1, ex)]
+                             -F10 * fh[idx_fh_F(iF, jF, kF    , ex)]
+                             +F18 * fh[idx_fh_F(iF, jF, kF + 1, ex)]
+                             -F6  * fh[idx_fh_F(iF, jF, kF + 2, ex)]
+                             +      fh[idx_fh_F(iF, jF, kF + 3, ex)]);
                     }
                 }
             }
         }
-
-        /* ---- KO dissipation (r=2, cof=16, sign=-) ---- */
-        if (eps > ZEO) {
-            const double cof = 16.0;
-            const double F4k = 4.0, F6k = 6.0;
-            const int i0_lo = (iminF+1>0)?iminF+1:0, j0_lo=(jminF+1>0)?jminF+1:0, k0_lo=(kminF+1>0)?kminF+1:0;
-            const int i0_hi=imaxF-3, j0_hi=jmaxF-3, k0_hi=kmaxF-3;
-            if (!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)) {
-                for (int k0=k0_lo;k0<=k0_hi;++k0) { const int kF=k0+1;
-                for (int j0=j0_lo;j0<=j0_hi;++j0) { const int jF=j0+1;
-                for (int i0=i0_lo;i0<=i0_hi;++i0) { const int iF=i0+1;
-                    const size_t p=idx_ex(i0,j0,k0,ex);
-                    const double Dx=((fh[idx_fh_F_ord2(iF-2,jF,kF,ex)]+fh[idx_fh_F_ord2(iF+2,jF,kF,ex)])-F4k*(fh[idx_fh_F_ord2(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord2(iF+1,jF,kF,ex)])+F6k*fh[idx_fh_F_ord2(iF,jF,kF,ex)])/dX;
-                    const double Dy=((fh[idx_fh_F_ord2(iF,jF-2,kF,ex)]+fh[idx_fh_F_ord2(iF,jF+2,kF,ex)])-F4k*(fh[idx_fh_F_ord2(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord2(iF,jF+1,kF,ex)])+F6k*fh[idx_fh_F_ord2(iF,jF,kF,ex)])/dY;
-                    const double Dz=((fh[idx_fh_F_ord2(iF,jF,kF-2,ex)]+fh[idx_fh_F_ord2(iF,jF,kF+2,ex)])-F4k*(fh[idx_fh_F_ord2(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord2(iF,jF,kF+1,ex)])+F6k*fh[idx_fh_F_ord2(iF,jF,kF,ex)])/dZ;
-                    f_rhs[p] -= (eps/cof)*(Dx+Dy+Dz);
-                }}}
-            }
-        }
-        free(fh);
-        return;
     }
-#elif (ghost_width == 3)
-    /* ---- 4th-order advection + r=3 KO (original code) ----------------- */
-    {
-        const int ord = 3;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -2;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -2;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -2;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        double *fh = (double*)malloc(nx*ny*nz*sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
+    // KO dissipation (same domain restriction as kodiss_c.C)
+    if (eps > ZEO) {
+        const int i0_lo = (iminF + 2 > 0) ? iminF + 2 : 0;
+        const int j0_lo = (jminF + 2 > 0) ? jminF + 2 : 0;
+        const int k0_lo = (kminF + 2 > 0) ? kminF + 2 : 0;
+        const int i0_hi = imaxF - 4; // inclusive
+        const int j0_hi = jmaxF - 4;
+        const int k0_hi = kmaxF - 4;
 
-        const double d12dx = ONE/F12/dX, d12dy = ONE/F12/dY, d12dz = ONE/F12/dZ;
+        if (!(i0_lo > i0_hi || j0_lo > j0_hi || k0_lo > k0_hi)) {
+            for (int k0 = k0_lo; k0 <= k0_hi; ++k0) {
+                const int kF = k0 + 1;
+                for (int j0 = j0_lo; j0 <= j0_hi; ++j0) {
+                    const int jF = j0 + 1;
+                    for (int i0 = i0_lo; i0 <= i0_hi; ++i0) {
+                        const int iF = i0 + 1;
+                        const size_t p = idx_ex(i0, j0, k0, ex);
 
-        /* ---- advection ---- */
-        for (int k0 = 0; k0 <= ex3-2; ++k0) {
-            const int kF = k0+1;
-            for (int j0 = 0; j0 <= ex2-2; ++j0) {
-                const int jF = j0+1;
-                for (int i0 = 0; i0 <= ex1-2; ++i0) {
-                    const int iF = i0+1;
-                    const size_t p = idx_ex(i0,j0,k0,ex);
+                        const double Dx_term =
+                            ((fh[idx_fh_F(iF - 3, jF, kF, ex)] + fh[idx_fh_F(iF + 3, jF, kF, ex)]) -
+                             SIX * (fh[idx_fh_F(iF - 2, jF, kF, ex)] + fh[idx_fh_F(iF + 2, jF, kF, ex)]) +
+                             FIT * (fh[idx_fh_F(iF - 1, jF, kF, ex)] + fh[idx_fh_F(iF + 1, jF, kF, ex)]) -
+                             TWT *  fh[idx_fh_F(iF,     jF, kF, ex)]) / dX;
 
-                    const double sfx = Sfx[p];
-                    if (sfx > ZEO) {
-                        if (i0 <= ex1-4)
-                            f_rhs[p] += sfx*d12dx*(-F3*fh[idx_fh_F(iF-1,jF,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF+1,jF,kF,ex)]-F6*fh[idx_fh_F(iF+2,jF,kF,ex)]+fh[idx_fh_F(iF+3,jF,kF,ex)]);
-                        else if (i0 <= ex1-3)
-                            f_rhs[p] += sfx*d12dx*(fh[idx_fh_F(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F(iF-1,jF,kF,ex)]+EIT*fh[idx_fh_F(iF+1,jF,kF,ex)]-fh[idx_fh_F(iF+2,jF,kF,ex)]);
-                        else if (i0 <= ex1-2)
-                            f_rhs[p] -= sfx*d12dx*(-F3*fh[idx_fh_F(iF+1,jF,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF-1,jF,kF,ex)]-F6*fh[idx_fh_F(iF-2,jF,kF,ex)]+fh[idx_fh_F(iF-3,jF,kF,ex)]);
-                    } else if (sfx < ZEO) {
-                        if ((i0-2) >= iminF)
-                            f_rhs[p] -= sfx*d12dx*(-F3*fh[idx_fh_F(iF+1,jF,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF-1,jF,kF,ex)]-F6*fh[idx_fh_F(iF-2,jF,kF,ex)]+fh[idx_fh_F(iF-3,jF,kF,ex)]);
-                        else if ((i0-1) >= iminF)
-                            f_rhs[p] += sfx*d12dx*(fh[idx_fh_F(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F(iF-1,jF,kF,ex)]+EIT*fh[idx_fh_F(iF+1,jF,kF,ex)]-fh[idx_fh_F(iF+2,jF,kF,ex)]);
-                        else if (i0 >= iminF)
-                            f_rhs[p] += sfx*d12dx*(-F3*fh[idx_fh_F(iF-1,jF,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF+1,jF,kF,ex)]-F6*fh[idx_fh_F(iF+2,jF,kF,ex)]+fh[idx_fh_F(iF+3,jF,kF,ex)]);
-                    }
-                    const double sfy = Sfy[p];
-                    if (sfy > ZEO) {
-                        if (j0<=ex2-4) f_rhs[p] += sfy*d12dy*(-F3*fh[idx_fh_F(iF,jF-1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF+1,kF,ex)]-F6*fh[idx_fh_F(iF,jF+2,kF,ex)]+fh[idx_fh_F(iF,jF+3,kF,ex)]);
-                        else if (j0<=ex2-3) f_rhs[p] += sfy*d12dy*(fh[idx_fh_F(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F(iF,jF+1,kF,ex)]-fh[idx_fh_F(iF,jF+2,kF,ex)]);
-                        else if (j0<=ex2-2) f_rhs[p] -= sfy*d12dy*(-F3*fh[idx_fh_F(iF,jF+1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF-1,kF,ex)]-F6*fh[idx_fh_F(iF,jF-2,kF,ex)]+fh[idx_fh_F(iF,jF-3,kF,ex)]);
-                    } else if (sfy < ZEO) {
-                        if ((j0-2)>=jminF) f_rhs[p] -= sfy*d12dy*(-F3*fh[idx_fh_F(iF,jF+1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF-1,kF,ex)]-F6*fh[idx_fh_F(iF,jF-2,kF,ex)]+fh[idx_fh_F(iF,jF-3,kF,ex)]);
-                        else if ((j0-1)>=jminF) f_rhs[p] += sfy*d12dy*(fh[idx_fh_F(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F(iF,jF+1,kF,ex)]-fh[idx_fh_F(iF,jF+2,kF,ex)]);
-                        else if (j0>=jminF) f_rhs[p] += sfy*d12dy*(-F3*fh[idx_fh_F(iF,jF-1,kF,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF+1,kF,ex)]-F6*fh[idx_fh_F(iF,jF+2,kF,ex)]+fh[idx_fh_F(iF,jF+3,kF,ex)]);
-                    }
-                    const double sfz = Sfz[p];
-                    if (sfz > ZEO) {
-                        if (k0<=ex3-4) f_rhs[p] += sfz*d12dz*(-F3*fh[idx_fh_F(iF,jF,kF-1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF,kF+1,ex)]-F6*fh[idx_fh_F(iF,jF,kF+2,ex)]+fh[idx_fh_F(iF,jF,kF+3,ex)]);
-                        else if (k0<=ex3-3) f_rhs[p] += sfz*d12dz*(fh[idx_fh_F(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F(iF,jF,kF+1,ex)]-fh[idx_fh_F(iF,jF,kF+2,ex)]);
-                        else if (k0<=ex3-2) f_rhs[p] -= sfz*d12dz*(-F3*fh[idx_fh_F(iF,jF,kF+1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF,kF-1,ex)]-F6*fh[idx_fh_F(iF,jF,kF-2,ex)]+fh[idx_fh_F(iF,jF,kF-3,ex)]);
-                    } else if (sfz < ZEO) {
-                        if ((k0-2)>=kminF) f_rhs[p] -= sfz*d12dz*(-F3*fh[idx_fh_F(iF,jF,kF+1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF,kF-1,ex)]-F6*fh[idx_fh_F(iF,jF,kF-2,ex)]+fh[idx_fh_F(iF,jF,kF-3,ex)]);
-                        else if ((k0-1)>=kminF) f_rhs[p] += sfz*d12dz*(fh[idx_fh_F(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F(iF,jF,kF+1,ex)]-fh[idx_fh_F(iF,jF,kF+2,ex)]);
-                        else if (k0>=kminF) f_rhs[p] += sfz*d12dz*(-F3*fh[idx_fh_F(iF,jF,kF-1,ex)]-F10*fh[idx_fh_F(iF,jF,kF,ex)]+F18*fh[idx_fh_F(iF,jF,kF+1,ex)]-F6*fh[idx_fh_F(iF,jF,kF+2,ex)]+fh[idx_fh_F(iF,jF,kF+3,ex)]);
+                        const double Dy_term =
+                            ((fh[idx_fh_F(iF, jF - 3, kF, ex)] + fh[idx_fh_F(iF, jF + 3, kF, ex)]) -
+                             SIX * (fh[idx_fh_F(iF, jF - 2, kF, ex)] + fh[idx_fh_F(iF, jF + 2, kF, ex)]) +
+                             FIT * (fh[idx_fh_F(iF, jF - 1, kF, ex)] + fh[idx_fh_F(iF, jF + 1, kF, ex)]) -
+                             TWT *  fh[idx_fh_F(iF, jF,     kF, ex)]) / dY;
+
+                        const double Dz_term =
+                            ((fh[idx_fh_F(iF, jF, kF - 3, ex)] + fh[idx_fh_F(iF, jF, kF + 3, ex)]) -
+                             SIX * (fh[idx_fh_F(iF, jF, kF - 2, ex)] + fh[idx_fh_F(iF, jF, kF + 2, ex)]) +
+                             FIT * (fh[idx_fh_F(iF, jF, kF - 1, ex)] + fh[idx_fh_F(iF, jF, kF + 1, ex)]) -
+                             TWT *  fh[idx_fh_F(iF, jF, kF,     ex)]) / dZ;
+
+                        f_rhs[p] += (eps / cof) * (Dx_term + Dy_term + Dz_term);
                     }
                 }
             }
         }
-
-        /* ---- KO dissipation (r=3, cof=64, sign=+) ---- */
-        if (eps > ZEO) {
-            const double cof = 64.0;
-            const double SIX = 6.0, FIT = 15.0, TWT = 20.0;
-            const int i0_lo=(iminF+2>0)?iminF+2:0, j0_lo=(jminF+2>0)?jminF+2:0, k0_lo=(kminF+2>0)?kminF+2:0;
-            const int i0_hi=imaxF-4, j0_hi=jmaxF-4, k0_hi=kmaxF-4;
-            if (!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)) {
-                for (int k0=k0_lo;k0<=k0_hi;++k0) { const int kF=k0+1;
-                for (int j0=j0_lo;j0<=j0_hi;++j0) { const int jF=j0+1;
-                for (int i0=i0_lo;i0<=i0_hi;++i0) { const int iF=i0+1;
-                    const size_t p=idx_ex(i0,j0,k0,ex);
-                    const double Dx=((fh[idx_fh_F(iF-3,jF,kF,ex)]+fh[idx_fh_F(iF+3,jF,kF,ex)])-SIX*(fh[idx_fh_F(iF-2,jF,kF,ex)]+fh[idx_fh_F(iF+2,jF,kF,ex)])+FIT*(fh[idx_fh_F(iF-1,jF,kF,ex)]+fh[idx_fh_F(iF+1,jF,kF,ex)])-TWT*fh[idx_fh_F(iF,jF,kF,ex)])/dX;
-                    const double Dy=((fh[idx_fh_F(iF,jF-3,kF,ex)]+fh[idx_fh_F(iF,jF+3,kF,ex)])-SIX*(fh[idx_fh_F(iF,jF-2,kF,ex)]+fh[idx_fh_F(iF,jF+2,kF,ex)])+FIT*(fh[idx_fh_F(iF,jF-1,kF,ex)]+fh[idx_fh_F(iF,jF+1,kF,ex)])-TWT*fh[idx_fh_F(iF,jF,kF,ex)])/dY;
-                    const double Dz=((fh[idx_fh_F(iF,jF,kF-3,ex)]+fh[idx_fh_F(iF,jF,kF+3,ex)])-SIX*(fh[idx_fh_F(iF,jF,kF-2,ex)]+fh[idx_fh_F(iF,jF,kF+2,ex)])+FIT*(fh[idx_fh_F(iF,jF,kF-1,ex)]+fh[idx_fh_F(iF,jF,kF+1,ex)])-TWT*fh[idx_fh_F(iF,jF,kF,ex)])/dZ;
-                    f_rhs[p] += (eps/cof)*(Dx+Dy+Dz);
-                }}}
-            }
-        }
-        free(fh);
-        return;
     }
-#elif (ghost_width == 4)
-    {
-        const int ord = 4;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -3;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -3;
 
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        double *fh = (double*)malloc(nx*ny*nz*sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d60dx=ONE/F60/dX, d60dy=ONE/F60/dY, d60dz=ONE/F60/dZ;
-        const double d12dx=ONE/F12/dX, d12dy=ONE/F12/dY, d12dz=ONE/F12/dZ;
-        const double d2dx=ONE/TWO/dX, d2dy=ONE/TWO/dY, d2dz=ONE/TWO/dZ;
-
-        /* ---- advection (6th-order lopsided) ---- */
-        for (int k0=0;k0<=ex3-2;++k0) { const int kF=k0+1;
-        for (int j0=0;j0<=ex2-2;++j0) { const int jF=j0+1;
-        for (int i0=0;i0<=ex1-2;++i0) { const int iF=i0+1;
-            const size_t p=idx_ex(i0,j0,k0,ex);
-            /* x */
-            const double sfx=Sfx[p];
-            if (sfx>ZEO) {
-                if (i0<=ex1-5&&(i0-1)>=iminF) f_rhs[p]+=sfx*d60dx*(+F2*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-F24*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F30*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]-fh[idx_fh_F_ord4(iF+4,jF,kF,ex)]);
-                else if (i0<=ex1-6&&i0>=iminF) f_rhs[p]+=sfx*d60dx*(-F10*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F100*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]+F50*fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]-F15*fh[idx_fh_F_ord4(iF+4,jF,kF,ex)]+F2*fh[idx_fh_F_ord4(iF+5,jF,kF,ex)]);
-                else if (i0<=ex1-4&&(i0-2)>=iminF) f_rhs[p]+=sfx*d60dx*(-fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-F45*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F9*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]);
-                else if (i0<=ex1-3&&(i0-1)>=iminF) f_rhs[p]+=sfx*d12dx*(fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]);
-                else if (i0<=ex1-2&&i0>=iminF) f_rhs[p]+=sfx*d2dx*(-fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]);
-            } else if (sfx<ZEO) {
-                if ((i0-4)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d60dx*(+F2*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]-F24*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]-F30*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]-fh[idx_fh_F_ord4(iF-4,jF,kF,ex)]);
-                else if ((i0-5)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d60dx*(-F10*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]-F100*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]+F50*fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]-F15*fh[idx_fh_F_ord4(iF-4,jF,kF,ex)]+F2*fh[idx_fh_F_ord4(iF-5,jF,kF,ex)]);
-                else if ((i0-3)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d60dx*(-fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-F45*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-F9*fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+3,jF,kF,ex)]);
-                else if ((i0-2)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d12dx*(fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord4(iF+2,jF,kF,ex)]);
-                else if ((i0-1)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d2dx*(-fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+1,jF,kF,ex)]);
-            }
-            /* y */
-            const double sfy=Sfy[p];
-            if (sfy>ZEO) {
-                if (j0<=ex2-5&&(j0-1)>=jminF) f_rhs[p]+=sfy*d60dy*(F2*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-F24*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F30*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]-fh[idx_fh_F_ord4(iF,jF+4,kF,ex)]);
-                else if (j0<=ex2-6&&j0>=jminF) f_rhs[p]+=sfy*d60dy*(-F10*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F100*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+F50*fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]-F15*fh[idx_fh_F_ord4(iF,jF+4,kF,ex)]+F2*fh[idx_fh_F_ord4(iF,jF+5,kF,ex)]);
-                else if (j0<=ex2-4&&(j0-2)>=jminF) f_rhs[p]+=sfy*d60dy*(-fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]);
-                else if (j0<=ex2-3&&(j0-1)>=jminF) f_rhs[p]+=sfy*d12dy*(fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]);
-                else if (j0<=ex2-2&&j0>=jminF) f_rhs[p]+=sfy*d2dy*(-fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]);
-            } else if (sfy<ZEO) {
-                if ((j0-4)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d60dy*(F2*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]-F24*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F30*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]-fh[idx_fh_F_ord4(iF,jF-4,kF,ex)]);
-                else if ((j0-5)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d60dy*(-F10*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]-F100*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]+F50*fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]-F15*fh[idx_fh_F_ord4(iF,jF-4,kF,ex)]+F2*fh[idx_fh_F_ord4(iF,jF-5,kF,ex)]);
-                else if ((j0-3)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d60dy*(-fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+3,kF,ex)]);
-                else if ((j0-2)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d12dy*(fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord4(iF,jF+2,kF,ex)]);
-                else if ((j0-1)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d2dy*(-fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+1,kF,ex)]);
-            }
-            /* z */
-            const double sfz=Sfz[p];
-            if (sfz>ZEO) {
-                if (k0<=ex3-5&&(k0-1)>=kminF) f_rhs[p]+=sfz*d60dz*(F2*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-F24*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F30*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]-fh[idx_fh_F_ord4(iF,jF,kF+4,ex)]);
-                else if (k0<=ex3-6&&k0>=kminF) f_rhs[p]+=sfz*d60dz*(-F10*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F100*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+F50*fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]-F15*fh[idx_fh_F_ord4(iF,jF,kF+4,ex)]+F2*fh[idx_fh_F_ord4(iF,jF,kF+5,ex)]);
-                else if (k0<=ex3-4&&(k0-2)>=kminF) f_rhs[p]+=sfz*d60dz*(-fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]);
-                else if (k0<=ex3-3&&(k0-1)>=kminF) f_rhs[p]+=sfz*d12dz*(fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]);
-                else if (k0<=ex3-2&&k0>=kminF) f_rhs[p]+=sfz*d2dz*(-fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]);
-            } else if (sfz<ZEO) {
-                if ((k0-4)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d60dz*(F2*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]-F24*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F35*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F80*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F30*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]-fh[idx_fh_F_ord4(iF,jF,kF-4,ex)]);
-                else if ((k0-5)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d60dz*(-F10*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F77*fh[idx_fh_F_ord4(iF,jF,kF,ex)]+F150*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]-F100*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]+F50*fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]-F15*fh[idx_fh_F_ord4(iF,jF,kF-4,ex)]+F2*fh[idx_fh_F_ord4(iF,jF,kF-5,ex)]);
-                else if ((k0-3)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d60dz*(-fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+3,ex)]);
-                else if ((k0-2)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d12dz*(fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord4(iF,jF,kF+2,ex)]);
-                else if ((k0-1)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d2dz*(-fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+1,ex)]);
-            }
-        }}}
-
-        /* ---- KO dissipation (r=4, cof=256, sign=-) ---- */
-        if (eps > ZEO) {
-            const double cof = 256.0;
-            const double F8k = 8.0, F28 = 28.0, F56 = 56.0, F70 = 70.0;
-            const int i0_lo=(iminF+3>0)?iminF+3:0, j0_lo=(jminF+3>0)?jminF+3:0, k0_lo=(kminF+3>0)?kminF+3:0;
-            const int i0_hi=imaxF-5, j0_hi=jmaxF-5, k0_hi=kmaxF-5;
-            if (!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)) {
-                for (int k0=k0_lo;k0<=k0_hi;++k0) { const int kF=k0+1;
-                for (int j0=j0_lo;j0<=j0_hi;++j0) { const int jF=j0+1;
-                for (int i0=i0_lo;i0<=i0_hi;++i0) { const int iF=i0+1;
-                    const size_t p=idx_ex(i0,j0,k0,ex);
-                    const double Dx=((fh[idx_fh_F_ord4(iF-4,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+4,jF,kF,ex)])-F8k*(fh[idx_fh_F_ord4(iF-3,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+3,jF,kF,ex)])+F28*(fh[idx_fh_F_ord4(iF-2,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+2,jF,kF,ex)])-F56*(fh[idx_fh_F_ord4(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord4(iF+1,jF,kF,ex)])+F70*fh[idx_fh_F_ord4(iF,jF,kF,ex)])/dX;
-                    const double Dy=((fh[idx_fh_F_ord4(iF,jF-4,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+4,kF,ex)])-F8k*(fh[idx_fh_F_ord4(iF,jF-3,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+3,kF,ex)])+F28*(fh[idx_fh_F_ord4(iF,jF-2,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+2,kF,ex)])-F56*(fh[idx_fh_F_ord4(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord4(iF,jF+1,kF,ex)])+F70*fh[idx_fh_F_ord4(iF,jF,kF,ex)])/dY;
-                    const double Dz=((fh[idx_fh_F_ord4(iF,jF,kF-4,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+4,ex)])-F8k*(fh[idx_fh_F_ord4(iF,jF,kF-3,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+3,ex)])+F28*(fh[idx_fh_F_ord4(iF,jF,kF-2,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+2,ex)])-F56*(fh[idx_fh_F_ord4(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord4(iF,jF,kF+1,ex)])+F70*fh[idx_fh_F_ord4(iF,jF,kF,ex)])/dZ;
-                    f_rhs[p] -= (eps/cof)*(Dx+Dy+Dz);
-                }}}
-            }
-        }
-        free(fh);
-        return;
-    }
-#elif (ghost_width == 5)
-    {
-        const int ord = 5;
-        int iminF = 1, jminF = 1, kminF = 1;
-        if (Symmetry > NO_SYMM && fabs(Z[0]) < dZ) kminF = -4;
-        if (Symmetry > EQ_SYMM && fabs(X[0]) < dX) iminF = -4;
-        if (Symmetry > EQ_SYMM && fabs(Y[0]) < dY) jminF = -4;
-
-        const size_t nx = (size_t)ex1 + ord;
-        const size_t ny = (size_t)ex2 + ord;
-        const size_t nz = (size_t)ex3 + ord;
-        double *fh = (double*)malloc(nx*ny*nz*sizeof(double));
-        if (!fh) return;
-        symmetry_bd(ord, ex, f, fh, SoA);
-
-        const double d840dx=ONE/F840/dX, d840dy=ONE/F840/dY, d840dz=ONE/F840/dZ;
-        const double d60dx=ONE/F60/dX, d60dy=ONE/F60/dY, d60dz=ONE/F60/dZ;
-        const double d12dx=ONE/F12/dX, d12dy=ONE/F12/dY, d12dz=ONE/F12/dZ;
-        const double d2dx=ONE/TWO/dX, d2dy=ONE/TWO/dY, d2dz=ONE/TWO/dZ;
-
-        /* ---- advection (8th-order lopsided) ---- */
-        for (int k0=0;k0<=ex3-2;++k0) { const int kF=k0+1;
-        for (int j0=0;j0<=ex2-2;++j0) { const int jF=j0+1;
-        for (int i0=0;i0<=ex1-2;++i0) { const int iF=i0+1;
-            const size_t p=idx_ex(i0,j0,k0,ex);
-            const double sfx=Sfx[p];
-            if (sfx>ZEO) {
-                if (i0<=ex1-6&&(i0-2)>=iminF) f_rhs[p]+=sfx*d840dx*(-F5*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F60*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-F420*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F420*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]+F140*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]-F30*fh[idx_fh_F_ord5(iF+4,jF,kF,ex)]+F3*fh[idx_fh_F_ord5(iF+5,jF,kF,ex)]);
-                else if (i0<=ex1-5&&(i0-3)>=iminF) f_rhs[p]+=sfx*d840dx*(+F3*fh[idx_fh_F_ord5(iF-4,jF,kF,ex)]-F32*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F168*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-F672*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+F672*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F168*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]+F32*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]-F3*fh[idx_fh_F_ord5(iF+4,jF,kF,ex)]);
-                else if (i0<=ex1-4&&(i0-2)>=iminF) f_rhs[p]+=sfx*d60dx*(-fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-F45*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F9*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]);
-                else if (i0<=ex1-3&&(i0-1)>=iminF) f_rhs[p]+=sfx*d12dx*(fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+EIT*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]);
-                else if (i0<=ex1-2&&i0>=iminF) f_rhs[p]+=sfx*d2dx*(-fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]);
-            } else if (sfx<ZEO) {
-                if ((i0-5)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d840dx*(-F5*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]+F60*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]-F420*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]-F420*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]+F140*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]-F30*fh[idx_fh_F_ord5(iF-4,jF,kF,ex)]+F3*fh[idx_fh_F_ord5(iF-5,jF,kF,ex)]);
-                else if ((i0-4)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d840dx*(+F3*fh[idx_fh_F_ord5(iF-4,jF,kF,ex)]-F32*fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F168*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-F672*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+F672*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F168*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]+F32*fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]-F3*fh[idx_fh_F_ord5(iF+4,jF,kF,ex)]);
-                else if ((i0-3)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d60dx*(-fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+F9*fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-F45*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+F45*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-F9*fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+3,jF,kF,ex)]);
-                else if ((i0-2)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d12dx*(fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+EIT*fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]-fh[idx_fh_F_ord5(iF+2,jF,kF,ex)]);
-                else if ((i0-1)>=iminF&&i0<=ex1-2) f_rhs[p]-=sfx*d2dx*(-fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+1,jF,kF,ex)]);
-            }
-            const double sfy=Sfy[p];
-            if (sfy>ZEO) {
-                if (j0<=ex2-6&&(j0-2)>=jminF) f_rhs[p]+=sfy*d840dy*(-F5*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F60*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+F140*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]-F30*fh[idx_fh_F_ord5(iF,jF+4,kF,ex)]+F3*fh[idx_fh_F_ord5(iF,jF+5,kF,ex)]);
-                else if (j0<=ex2-5&&(j0-3)>=jminF) f_rhs[p]+=sfy*d840dy*(+F3*fh[idx_fh_F_ord5(iF,jF-4,kF,ex)]-F32*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F168*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F672*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F672*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F168*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+F32*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]-F3*fh[idx_fh_F_ord5(iF,jF+4,kF,ex)]);
-                else if (j0<=ex2-4&&(j0-2)>=jminF) f_rhs[p]+=sfy*d60dy*(-fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]);
-                else if (j0<=ex2-3&&(j0-1)>=jminF) f_rhs[p]+=sfy*d12dy*(fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]);
-                else if (j0<=ex2-2&&j0>=jminF) f_rhs[p]+=sfy*d2dy*(-fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]);
-            } else if (sfy<ZEO) {
-                if ((j0-5)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d840dy*(-F5*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]+F60*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]-F420*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]+F140*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]-F30*fh[idx_fh_F_ord5(iF,jF-4,kF,ex)]+F3*fh[idx_fh_F_ord5(iF,jF-5,kF,ex)]);
-                else if ((j0-4)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d840dy*(+F3*fh[idx_fh_F_ord5(iF,jF-4,kF,ex)]-F32*fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F168*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F672*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F672*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F168*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+F32*fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]-F3*fh[idx_fh_F_ord5(iF,jF+4,kF,ex)]);
-                else if ((j0-3)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d60dy*(-fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+F9*fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-F45*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+F45*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-F9*fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+3,kF,ex)]);
-                else if ((j0-2)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d12dy*(fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]-fh[idx_fh_F_ord5(iF,jF+2,kF,ex)]);
-                else if ((j0-1)>=jminF&&j0<=ex2-2) f_rhs[p]-=sfy*d2dy*(-fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+1,kF,ex)]);
-            }
-            const double sfz=Sfz[p];
-            if (sfz>ZEO) {
-                if (k0<=ex3-6&&(k0-2)>=kminF) f_rhs[p]+=sfz*d840dz*(-F5*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F60*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+F140*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]-F30*fh[idx_fh_F_ord5(iF,jF,kF+4,ex)]+F3*fh[idx_fh_F_ord5(iF,jF,kF+5,ex)]);
-                else if (k0<=ex3-5&&(k0-3)>=kminF) f_rhs[p]+=sfz*d840dz*(+F3*fh[idx_fh_F_ord5(iF,jF,kF-4,ex)]-F32*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F168*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F672*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F672*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F168*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+F32*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]-F3*fh[idx_fh_F_ord5(iF,jF,kF+4,ex)]);
-                else if (k0<=ex3-4&&(k0-2)>=kminF) f_rhs[p]+=sfz*d60dz*(-fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]);
-                else if (k0<=ex3-3&&(k0-1)>=kminF) f_rhs[p]+=sfz*d12dz*(fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]);
-                else if (k0<=ex3-2&&k0>=kminF) f_rhs[p]+=sfz*d2dz*(-fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]);
-            } else if (sfz<ZEO) {
-                if ((k0-5)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d840dz*(-F5*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]+F60*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F378*fh[idx_fh_F_ord5(iF,jF,kF,ex)]+F1050*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]-F420*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]+F140*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]-F30*fh[idx_fh_F_ord5(iF,jF,kF-4,ex)]+F3*fh[idx_fh_F_ord5(iF,jF,kF-5,ex)]);
-                else if ((k0-4)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d840dz*(+F3*fh[idx_fh_F_ord5(iF,jF,kF-4,ex)]-F32*fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F168*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F672*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F672*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F168*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+F32*fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]-F3*fh[idx_fh_F_ord5(iF,jF,kF+4,ex)]);
-                else if ((k0-3)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d60dz*(-fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+F9*fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-F45*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+F45*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-F9*fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+3,ex)]);
-                else if ((k0-2)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d12dz*(fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]-EIT*fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+EIT*fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]-fh[idx_fh_F_ord5(iF,jF,kF+2,ex)]);
-                else if ((k0-1)>=kminF&&k0<=ex3-2) f_rhs[p]-=sfz*d2dz*(-fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+1,ex)]);
-            }
-        }}}
-
-        /* ---- KO dissipation (r=5, cof=1024, sign=+) ---- */
-        if (eps > ZEO) {
-            const double cof = 1024.0;
-            const double F10k=10.0, F45k=45.0, F120=120.0, F210=210.0, F252=252.0;
-            const int i0_lo=(iminF+4>0)?iminF+4:0, j0_lo=(jminF+4>0)?jminF+4:0, k0_lo=(kminF+4>0)?kminF+4:0;
-            const int i0_hi=imaxF-6, j0_hi=jmaxF-6, k0_hi=kmaxF-6;
-            if (!(i0_lo>i0_hi||j0_lo>j0_hi||k0_lo>k0_hi)) {
-                for (int k0=k0_lo;k0<=k0_hi;++k0) { const int kF=k0+1;
-                for (int j0=j0_lo;j0<=j0_hi;++j0) { const int jF=j0+1;
-                for (int i0=i0_lo;i0<=i0_hi;++i0) { const int iF=i0+1;
-                    const size_t p=idx_ex(i0,j0,k0,ex);
-                    const double Dx=((fh[idx_fh_F_ord5(iF-5,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+5,jF,kF,ex)])-F10k*(fh[idx_fh_F_ord5(iF-4,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+4,jF,kF,ex)])+F45k*(fh[idx_fh_F_ord5(iF-3,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+3,jF,kF,ex)])-F120*(fh[idx_fh_F_ord5(iF-2,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+2,jF,kF,ex)])+F210*(fh[idx_fh_F_ord5(iF-1,jF,kF,ex)]+fh[idx_fh_F_ord5(iF+1,jF,kF,ex)])-F252*fh[idx_fh_F_ord5(iF,jF,kF,ex)])/dX;
-                    const double Dy=((fh[idx_fh_F_ord5(iF,jF-5,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+5,kF,ex)])-F10k*(fh[idx_fh_F_ord5(iF,jF-4,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+4,kF,ex)])+F45k*(fh[idx_fh_F_ord5(iF,jF-3,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+3,kF,ex)])-F120*(fh[idx_fh_F_ord5(iF,jF-2,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+2,kF,ex)])+F210*(fh[idx_fh_F_ord5(iF,jF-1,kF,ex)]+fh[idx_fh_F_ord5(iF,jF+1,kF,ex)])-F252*fh[idx_fh_F_ord5(iF,jF,kF,ex)])/dY;
-                    const double Dz=((fh[idx_fh_F_ord5(iF,jF,kF-5,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+5,ex)])-F10k*(fh[idx_fh_F_ord5(iF,jF,kF-4,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+4,ex)])+F45k*(fh[idx_fh_F_ord5(iF,jF,kF-3,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+3,ex)])-F120*(fh[idx_fh_F_ord5(iF,jF,kF-2,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+2,ex)])+F210*(fh[idx_fh_F_ord5(iF,jF,kF-1,ex)]+fh[idx_fh_F_ord5(iF,jF,kF+1,ex)])-F252*fh[idx_fh_F_ord5(iF,jF,kF,ex)])/dZ;
-                    f_rhs[p] += (eps/cof)*(Dx+Dy+Dz);
-                }}}
-            }
-        }
-        free(fh);
-        return;
-    }
-#else
-#error "lopsided_kodis_c.C: unsupported ghost_width (must be 2, 3, 4, or 5)"
-#endif
+    free(fh);
 }

AMSS_NCKU_source/macrodef.h

@@ -29,16 +29,6 @@
 
 #define REGLEV 0
 
-#define BSSN_FINE_TIMING 0
-
-#define BSSN_FINE_TIMING_EVERY 1
-
-#define BSSN_FINE_TIMING_TOPN 8
-
-#define BSSN_KERNEL_FINE_TIMING 0
-
-#define BSSN_ENABLE_STDIN_ABORT_POLL 0
-
 //#define USE_GPU
 
 //#define CHECKDETAIL
@@ -98,21 +88,6 @@
 //     0: for every level;
 //     1: for all
 //
-// define BSSN_FINE_TIMING
-//     enable fine-grained per-timestep timing monitor
-//
-// define BSSN_FINE_TIMING_EVERY
-//     report timing every N coarse timesteps
-//
-// define BSSN_FINE_TIMING_TOPN
-//     number of hottest timing buckets shown in stdout
-//
-// define BSSN_KERNEL_FINE_TIMING
-//     enable split timing inside compute_rhs_bssn
-//
-// define BSSN_ENABLE_STDIN_ABORT_POLL
-//     poll stdin and broadcast abort flag every coarse step
-//
 // define USE_GPU
 //     use gpu or not
 //
@@ -167,3 +142,4 @@
 #define TINY 1e-10
 
 #endif   /* MICRODEF_H */
+

AMSS_NCKU_source/makefile

@@ -1,189 +1,93 @@
-
-
+
+
 include makefile.inc
 
--include AMSS_NCKU_build.mk
-
-ABE_TYPE ?= $(shell awk '/^[[:space:]]*\#define[[:space:]]+ABEtype/ {print $$3; exit}' macrodef.h 2>/dev/null)
-
-ifeq ($(USE_TRANSFER_CACHE),auto)
-ifeq ($(ABE_TYPE),0)
-EFFECTIVE_USE_TRANSFER_CACHE = 1
-else
-EFFECTIVE_USE_TRANSFER_CACHE = 0
-endif
-else
-EFFECTIVE_USE_TRANSFER_CACHE = $(USE_TRANSFER_CACHE)
-endif
-
-ifeq ($(USE_CXX_ESCALAR_KERNEL),1)
-ifeq ($(ABE_TYPE),1)
-EFFECTIVE_USE_CXX_ESCALAR_KERNEL = 1
-else
-EFFECTIVE_USE_CXX_ESCALAR_KERNEL = 0
-endif
-else
-EFFECTIVE_USE_CXX_ESCALAR_KERNEL = 0
-endif
-
-ifeq ($(EFFECTIVE_USE_CXX_ESCALAR_KERNEL),1)
-ifeq ($(USE_CXX_KERNELS),0)
-$(error USE_CXX_ESCALAR_KERNEL=1 requires USE_CXX_KERNELS=1 because bssn_escalar_rhs_c.C reuses the C BSSN kernel)
-endif
-endif
-
-ifeq ($(USE_CXX_EM_KERNEL),1)
-ifeq ($(ABE_TYPE),3)
-EFFECTIVE_USE_CXX_EM_KERNEL = 1
-else
-EFFECTIVE_USE_CXX_EM_KERNEL = 0
-endif
-else
-EFFECTIVE_USE_CXX_EM_KERNEL = 0
-endif
-
-ifeq ($(EFFECTIVE_USE_CXX_EM_KERNEL),1)
-ifeq ($(USE_CXX_KERNELS),0)
-$(error USE_CXX_EM_KERNEL=1 requires USE_CXX_KERNELS=1 because bssn_em_rhs_c.C reuses the C BSSN kernel)
-endif
-endif
-
-EM_KERNEL_FLAG = -DBSSN_USE_EM_C_KERNEL=$(EFFECTIVE_USE_CXX_EM_KERNEL)
-
 ## polint(ordn=6) kernel selector:
 ##   1 (default): barycentric fast path
 ##   0          : fallback to Neville path
 POLINT6_USE_BARY ?= 1
 POLINT6_FLAG = -DPOLINT6_USE_BARYCENTRIC=$(POLINT6_USE_BARY)
-FMISC_SAFE_FLAG = -DUSE_FMISC_SAFE_MODE=$(USE_FMISC_SAFE_MODE)
-TRANSFER_CACHE_FLAG = -DBSSN_USE_TRANSFER_CACHE=$(EFFECTIVE_USE_TRANSFER_CACHE)
-ESCALAR_KERNEL_FLAG = -DBSSN_USE_ESCALAR_C_KERNEL=$(EFFECTIVE_USE_CXX_ESCALAR_KERNEL)
 
 ## ABE build flags selected by PGO_MODE (set in makefile.inc, default: opt)
 ##   make                        -> opt  (PGO-guided, maximum performance)
 ##   make PGO_MODE=instrument    -> instrument (Phase 1: collect fresh profile data)
 PROFDATA = /home/$(shell whoami)/AMSS-NCKU/pgo_profile/default.profdata
-
-ifeq ($(PGO_MODE),instrument)
-## Phase 1: instrumentation — omit -ipo/-fp-model fast=2 for faster build and numerical stability
-CXXAPPFLAGS = -O3 -march=x86-64-v4 -fma -fprofile-instr-generate -ipo \
-              -Dfortran3 -Dnewc -I${MKLROOT}/include $(INTERP_LB_FLAGS) \
-              $(FMISC_SAFE_FLAG) \
-              $(TRANSFER_CACHE_FLAG) $(ESCALAR_KERNEL_FLAG) $(EM_KERNEL_FLAG)
-f90appflags = -O3 -march=x86-64-v4 -fma -fprofile-instr-generate -ipo \
-              -align array64byte -fpp -I${MKLROOT}/include $(POLINT6_FLAG) \
-              $(FMISC_SAFE_FLAG)
+
+ifeq ($(PGO_MODE),instrument)
+## Phase 1: instrumentation — omit -ipo/-fp-model fast=2 for faster build and numerical stability
+CXXAPPFLAGS = -O3 -xHost -fma -fprofile-instr-generate -ipo \
+              -Dfortran3 -Dnewc -I${MKLROOT}/include $(INTERP_LB_FLAGS)
+f90appflags = -O3 -xHost -fma -fprofile-instr-generate -ipo \
+              -align array64byte -fpp -I${MKLROOT}/include $(POLINT6_FLAG)
 else
 ## opt (default): maximum performance with PGO profile data -fprofile-instr-use=$(PROFDATA) \
 ## PGO has been turned off, now tested and found to be negative optimization
 ## INTERP_LB_FLAGS has been turned off too, now tested and found to be negative optimization
 
 
-CXXAPPFLAGS = -O3 -march=x86-64-v4 -fp-model fast=2 -fma -ipo \
-              -Dfortran3 -Dnewc -I${MKLROOT}/include $(INTERP_LB_FLAGS) \
-              $(FMISC_SAFE_FLAG) \
-              $(TRANSFER_CACHE_FLAG) $(ESCALAR_KERNEL_FLAG) $(EM_KERNEL_FLAG)
-f90appflags = -O3 -march=x86-64-v4 -fp-model fast=2 -fma -ipo \
-              -align array64byte -fpp -I${MKLROOT}/include $(POLINT6_FLAG) \
-              $(FMISC_SAFE_FLAG)
+CXXAPPFLAGS = -O3 -xHost -fp-model fast=2 -fma -ipo \
+              -Dfortran3 -Dnewc -I${MKLROOT}/include $(INTERP_LB_FLAGS)
+f90appflags = -O3 -xHost -fp-model fast=2 -fma -ipo \
+              -align array64byte -fpp -I${MKLROOT}/include $(POLINT6_FLAG)
 endif
 
 .SUFFIXES: .o .f90 .C .for .cu
-
-.f90.o:
-	$(f90) $(f90appflags) -c $< -o $@
-
-.C.o:
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-# ShellPatch.C uses OpenMP for setupintintstuff search loops
-ShellPatch.o: ShellPatch.C
-	${CXX} $(CXXAPPFLAGS) $(OMP_FLAG) -c $< $(filein) -o $@
-
-.for.o:
-	$(f77) -c $< -o $@
-
-.cu.o:
-	$(Cu) $(CUDA_APP_FLAGS) -c $< -o $@ $(CUDA_LIB_PATH)
-
-# C rewrite of BSSN RHS kernel and helpers
-bssn_rhs_c.o: bssn_rhs_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-fderivs_c.o: fderivs_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-fdderivs_c.o: fdderivs_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-kodiss_c.o: kodiss_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
+
+.f90.o:
+	$(f90) $(f90appflags) -c $< -o $@
+
+.C.o:
+	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
+
+.for.o:
+	$(f77) -c $< -o $@
+
+.cu.o:
+	$(Cu) $(CUDA_APP_FLAGS) -c $< -o $@ $(CUDA_LIB_PATH)
+
+# C rewrite of BSSN RHS kernel and helpers
+bssn_rhs_c.o: bssn_rhs_c.C
+	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
+
+fderivs_c.o: fderivs_c.C
+	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
+
+fdderivs_c.o: fdderivs_c.C
+	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
+
+kodiss_c.o: kodiss_c.C
+	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
+
 lopsided_c.o: lopsided_c.C
 	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
 
 lopsided_kodis_c.o: lopsided_kodis_c.C
 	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
 
-# C rewrite of shell-patch derivative kernels
-fderivs_sh_c.o: fderivs_sh_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-fdderivs_sh_c.o: fdderivs_sh_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-fderivs_shc_c.o: fderivs_shc_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-fdderivs_shc_c.o: fdderivs_shc_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-kodiss_sh_c.o: kodiss_sh_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-
-bssn_em_rhs_c.o: bssn_em_rhs_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-z4c_rhs_c.o: z4c_rhs_c.C
-	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
 #interp_lb_profile.o: interp_lb_profile.C interp_lb_profile.h
 #	${CXX} $(CXXAPPFLAGS) -c $< $(filein) -o $@
-
-## TwoPunctureABE uses fixed optimal flags with its own PGO profile, independent of CXXAPPFLAGS
-TP_PROFDATA = /home/$(shell whoami)/AMSS-NCKU/pgo_profile/TwoPunctureABE.profdata
-TP_OPTFLAGS = -O3 -march=x86-64-v4 -fp-model fast=2 -fma -ipo \
-              -fprofile-instr-use=$(TP_PROFDATA) \
-              -Dfortran3 -Dnewc -I${MKLROOT}/include
-
-TwoPunctures.o: TwoPunctures.C
-	${CXX} $(TP_OPTFLAGS) -qopenmp -c $< -o $@
-
-TwoPunctureABE.o: TwoPunctureABE.C
-	${CXX} $(TP_OPTFLAGS) -qopenmp -c $< -o $@
-
-# Input files
-
-## Kernel implementation switch (set USE_CXX_KERNELS=0 to fall back to Fortran)
+
+## TwoPunctureABE uses fixed optimal flags with its own PGO profile, independent of CXXAPPFLAGS
+TP_PROFDATA = /home/$(shell whoami)/AMSS-NCKU/pgo_profile/TwoPunctureABE.profdata
+TP_OPTFLAGS = -O3 -xHost -fp-model fast=2 -fma -ipo \
+              -fprofile-instr-use=$(TP_PROFDATA) \
+              -Dfortran3 -Dnewc -I${MKLROOT}/include
+
+TwoPunctures.o: TwoPunctures.C
+	${CXX} $(TP_OPTFLAGS) -qopenmp -c $< -o $@
+
+TwoPunctureABE.o: TwoPunctureABE.C
+	${CXX} $(TP_OPTFLAGS) -qopenmp -c $< -o $@
+
+# Input files
+
+## Kernel implementation switch (set USE_CXX_KERNELS=0 to fall back to Fortran)
 ifeq ($(USE_CXX_KERNELS),0)
 # Fortran mode: no C rewrite files; bssn_rhs.o is included via F90FILES below
 CFILES =
 else
-# C++ mode (default): C rewrite of bssn/bssn-escalar rhs and helper kernels
+# C++ mode (default): C rewrite of bssn_rhs and helper kernels
 CFILES = bssn_rhs_c.o fderivs_c.o fdderivs_c.o kodiss_c.o lopsided_c.o lopsided_kodis_c.o
-ifeq ($(EFFECTIVE_USE_CXX_ESCALAR_KERNEL),1)
-CFILES += bssn_escalar_rhs_c.o
-endif
-ifeq ($(EFFECTIVE_USE_CXX_EM_KERNEL),1)
-CFILES += bssn_em_rhs_c.o
-endif
-endif
-
-ifeq ($(USE_CXX_Z4C_KERNELS),1)
-CFILES += z4c_rhs_c.o
-Z4C_F90_OBJ =
-else
-Z4C_F90_OBJ = Z4c_rhs.o
 endif
 
 ## RK4 kernel switch (independent from USE_CXX_KERNELS)
@@ -193,106 +97,94 @@ RK4_F90_OBJ =
 else
 RK4_F90_OBJ = rungekutta4_rout.o
 endif
-
-## Shell-patch derivative kernel switch (independent from USE_CXX_KERNELS)
-##   1 : use C++ rewrite of shell derivative functions (experimental)
-##   0 : use original Fortran diff_new_sh.o and kodiss_sh.o (default)
-USE_CXX_SHELL_KERNELS ?= 0
-ifeq ($(USE_CXX_SHELL_KERNELS),1)
-CFILES += fderivs_sh_c.o fdderivs_sh_c.o fderivs_shc_c.o fdderivs_shc_c.o kodiss_sh_c.o
-SH_F90_OBJ =
-else
-SH_F90_OBJ = diff_new_sh.o kodiss_sh.o point_diff_new_sh.o
-endif
-
-C++FILES = ABE.o Ansorg.o Block.o misc.o monitor.o Parallel.o MPatch.o var.o\
+
+C++FILES = ABE.o Ansorg.o Block.o misc.o monitor.o Parallel.o MPatch.o var.o\
+           cgh.o bssn_class.o surface_integral.o ShellPatch.o\
+	   bssnEScalar_class.o perf.o Z4c_class.o NullShellPatch.o\
+	   bssnEM_class.o cpbc_util.o z4c_rhs_point.o checkpoint.o\
+           Parallel_bam.o scalar_class.o transpbh.o NullShellPatch2.o\
+	   NullShellPatch2_Evo.o writefile_f.o interp_lb_profile.o
+	   
+C++FILES_GPU = ABE.o Ansorg.o Block.o misc.o monitor.o Parallel.o MPatch.o var.o\
            cgh.o bssn_class.o surface_integral.o ShellPatch.o\
 	   bssnEScalar_class.o perf.o Z4c_class.o NullShellPatch.o\
 	   bssnEM_class.o cpbc_util.o z4c_rhs_point.o checkpoint.o\
            Parallel_bam.o scalar_class.o transpbh.o NullShellPatch2.o\
-	   NullShellPatch2_Evo.o writefile_f.o interp_lb_profile.o
-	   
-C++FILES_GPU = ABE.o Ansorg.o Block.o misc.o monitor.o Parallel.o MPatch.o var.o\
-           cgh.o surface_integral.o ShellPatch.o\
-	   bssnEScalar_class.o perf.o Z4c_class.o NullShellPatch.o\
-	   bssnEM_class.o cpbc_util.o z4c_rhs_point.o checkpoint.o\
-           Parallel_bam.o scalar_class.o transpbh.o NullShellPatch2.o\
-	   NullShellPatch2_Evo.o \
-	   bssn_gpu_class.o bssn_step_gpu.o bssn_macro.o writefile_f.o
-
+	   NullShellPatch2_Evo.o bssn_cuda_step.o writefile_f.o
+
 F90FILES_BASE = enforce_algebra.o fmisc.o initial_puncture.o prolongrestrict.o\
 	   prolongrestrict_cell.o prolongrestrict_vertex.o\
-	   $(RK4_F90_OBJ) diff_new.o kodiss.o\
-	   lopsidediff.o sommerfeld_rout.o getnp4.o $(SH_F90_OBJ)\
+	   $(RK4_F90_OBJ) diff_new.o kodiss.o kodiss_sh.o\
+	   lopsidediff.o sommerfeld_rout.o getnp4.o diff_new_sh.o\
 	   shellfunctions.o bssn_rhs_ss.o Set_Rho_ADM.o\
            getnp4EScalar.o bssnEScalar_rhs.o bssn_constraint.o ricci_gamma.o\
-           fadmquantites_bssn.o $(Z4C_F90_OBJ) Z4c_rhs_ss.o\
-	   cpbc.o getnp4old.o NullEvol.o initial_null.o initial_maxwell.o\
-	   getnpem2.o empart.o NullNews.o fourdcurvature.o\
-	   bssn2adm.o adm_constraint.o adm_ricci_gamma.o\
-	   scalar_rhs.o initial_scalar.o NullEvol2.o initial_null2.o\
-	   NullNews2.o tool_f.o
-
-ifeq ($(USE_CXX_KERNELS),0)
-# Fortran mode: include original bssn_rhs.o
-F90FILES = $(F90FILES_BASE) bssn_rhs.o
-else
-# C++ mode (default): bssn_rhs.o replaced by C++ kernel
-F90FILES = $(F90FILES_BASE)
-endif
-
-F77FILES = zbesh.o
-
-AHFDOBJS = expansion.o expansion_Jacobian.o patch.o coords.o patch_info.o patch_interp.o patch_system.o \
-tgrid.o fd_grid.o ghost_zone.o array.o round.o norm.o fuzzy.o error_exit.o miscfp.o \
-linear_map.o cpm_map.o BH_diagnostics.o setup.o horizon_sequence.o find_horizons.o \
-initial_guess.o Newton.o Jacobian.o ilucg.o IntPnts0.o IntPnts.o
-
-TwoPunctureFILES = TwoPunctureABE.o TwoPunctures.o 
-
-CUDAFILES = bssn_gpu.o bssn_gpu_rhs_ss.o
-
-# file dependences
-$(C++FILES) $(C++FILES_GPU) $(F90FILES) $(CFILES) $(AHFDOBJS) $(CUDAFILES): macrodef.fh
-
-$(C++FILES): Block.h enforce_algebra.h fmisc.h initial_puncture.h macrodef.h\
-	     misc.h monitor.h MyList.h Parallel.h MPatch.h prolongrestrict.h\
-	     rungekutta4_rout.h var.h bssn_class.h bssn_rhs.h sommerfeld_rout.h\
-	     cgh.h surface_integral.h ShellPatch.h shellfunctions.h perf.h\
-             fadmquantites_bssn.h cpbc.h getnp4.h initial_null.h NullEvol.h\
-	     NullShellPatch.h initial_maxwell.h bssnEM_class.h getnpem2.h\
-	     empart.h NullNews.h kodiss.h Parallel_bam.h ricci_gamma.h\
-             initial_null2.h NullShellPatch2.h 
-             
-$(C++FILES_GPU): Block.h enforce_algebra.h fmisc.h initial_puncture.h macrodef.h\
-	     misc.h monitor.h MyList.h Parallel.h MPatch.h prolongrestrict.h\
-	     rungekutta4_rout.h var.h bssn_rhs.h sommerfeld_rout.h\
-	     cgh.h surface_integral.h ShellPatch.h shellfunctions.h perf.h\
-             fadmquantites_bssn.h cpbc.h getnp4.h initial_null.h NullEvol.h\
-	     NullShellPatch.h initial_maxwell.h bssnEM_class.h getnpem2.h\
-	     empart.h NullNews.h kodiss.h Parallel_bam.h ricci_gamma.h\
-             initial_null2.h NullShellPatch2.h \
-             bssn_gpu_class.h bssn_macro.h
-             
-$(AHFDOBJS): cctk.h cctk_Config.h cctk_Types.h cctk_Constants.h myglobal.h
-
-$(C++FILES) $(C++FILES_GPU) $(CFILES) $(AHFDOBJS) $(CUDAFILES): macrodef.h
-
-TwoPunctureFILES: TwoPunctures.h
-
-$(CUDAFILES): bssn_gpu.h gpu_mem.h gpu_rhsSS_mem.h
-
-misc.o : zbesh.o
-
-# projects
-ABE: $(C++FILES) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS)
-	$(CLINKER) $(CXXAPPFLAGS) -o $@ $(C++FILES) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS) $(LDLIBS)
-	
-ABEGPU: $(C++FILES_GPU) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS) $(CUDAFILES)
-	$(CLINKER) $(CXXAPPFLAGS) -o $@ $(C++FILES_GPU) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS) $(CUDAFILES) $(LDLIBS)
-
-TwoPunctureABE: $(TwoPunctureFILES)
-	$(CLINKER) $(TP_OPTFLAGS) -qopenmp -o $@ $(TwoPunctureFILES) $(LDLIBS)
-
-clean:
-	rm *.o ABE ABEGPU TwoPunctureABE make.log -f
+           fadmquantites_bssn.o Z4c_rhs.o Z4c_rhs_ss.o point_diff_new_sh.o\
+	   cpbc.o getnp4old.o NullEvol.o initial_null.o initial_maxwell.o\
+	   getnpem2.o empart.o NullNews.o fourdcurvature.o\
+	   bssn2adm.o adm_constraint.o adm_ricci_gamma.o\
+	   scalar_rhs.o initial_scalar.o NullEvol2.o initial_null2.o\
+	   NullNews2.o tool_f.o
+
+ifeq ($(USE_CXX_KERNELS),0)
+# Fortran mode: include original bssn_rhs.o
+F90FILES = $(F90FILES_BASE) bssn_rhs.o
+else
+# C++ mode (default): bssn_rhs.o replaced by C++ kernel
+F90FILES = $(F90FILES_BASE)
+endif
+
+F77FILES = zbesh.o
+
+AHFDOBJS = expansion.o expansion_Jacobian.o patch.o coords.o patch_info.o patch_interp.o patch_system.o \
+tgrid.o fd_grid.o ghost_zone.o array.o round.o norm.o fuzzy.o error_exit.o miscfp.o \
+linear_map.o cpm_map.o BH_diagnostics.o setup.o horizon_sequence.o find_horizons.o \
+initial_guess.o Newton.o Jacobian.o ilucg.o IntPnts0.o IntPnts.o
+
+TwoPunctureFILES = TwoPunctureABE.o TwoPunctures.o 
+
+CUDAFILES = bssn_gpu.o bssn_cuda_ops.o
+
+# file dependences
+$(C++FILES) $(C++FILES_GPU) $(F90FILES) $(CFILES) $(AHFDOBJS) $(CUDAFILES): macrodef.fh
+
+$(C++FILES): Block.h enforce_algebra.h fmisc.h initial_puncture.h macrodef.h\
+	     misc.h monitor.h MyList.h Parallel.h MPatch.h prolongrestrict.h\
+	     rungekutta4_rout.h var.h bssn_class.h bssn_rhs.h sommerfeld_rout.h\
+	     cgh.h surface_integral.h ShellPatch.h shellfunctions.h perf.h\
+             fadmquantites_bssn.h cpbc.h getnp4.h initial_null.h NullEvol.h\
+	     NullShellPatch.h initial_maxwell.h bssnEM_class.h getnpem2.h\
+	     empart.h NullNews.h kodiss.h Parallel_bam.h ricci_gamma.h\
+             initial_null2.h NullShellPatch2.h 
+             
+$(C++FILES_GPU): Block.h enforce_algebra.h fmisc.h initial_puncture.h macrodef.h\
+	     misc.h monitor.h MyList.h Parallel.h MPatch.h prolongrestrict.h\
+	     rungekutta4_rout.h var.h bssn_rhs.h sommerfeld_rout.h\
+	     cgh.h surface_integral.h ShellPatch.h shellfunctions.h perf.h\
+             fadmquantites_bssn.h cpbc.h getnp4.h initial_null.h NullEvol.h\
+	     NullShellPatch.h initial_maxwell.h bssnEM_class.h getnpem2.h\
+	     empart.h NullNews.h kodiss.h Parallel_bam.h ricci_gamma.h\
+             initial_null2.h NullShellPatch2.h \
+             bssn_gpu_class.h bssn_macro.h
+             
+$(AHFDOBJS): cctk.h cctk_Config.h cctk_Types.h cctk_Constants.h myglobal.h
+
+$(C++FILES) $(C++FILES_GPU) $(CFILES) $(AHFDOBJS) $(CUDAFILES): macrodef.h
+
+TwoPunctureFILES: TwoPunctures.h
+
+$(CUDAFILES): bssn_gpu.h gpu_mem.h gpu_rhsSS_mem.h
+
+misc.o : zbesh.o
+
+# projects
+ABE: $(C++FILES) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS)
+	$(CLINKER) $(CXXAPPFLAGS) -o $@ $(C++FILES) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS) $(LDLIBS)
+	
+ABEGPU: $(C++FILES_GPU) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS) $(CUDAFILES)
+	$(CLINKER) $(CXXAPPFLAGS) -o $@ $(C++FILES_GPU) $(CFILES) $(F90FILES) $(F77FILES) $(AHFDOBJS) $(CUDAFILES) $(LDLIBS)
+
+TwoPunctureABE: $(TwoPunctureFILES)
+	$(CLINKER) $(TP_OPTFLAGS) -qopenmp -o $@ $(TwoPunctureFILES) $(LDLIBS)
+
+clean:
+	rm *.o ABE ABEGPU TwoPunctureABE make.log -f

AMSS_NCKU_source/makefile.inc

@@ -9,6 +9,7 @@ filein  = -I/usr/include/ -I${MKLROOT}/include
 ## Using sequential MKL (OpenMP disabled for better single-threaded performance)
 ## Added -lifcore for Intel Fortran runtime and -limf for Intel math library
 LDLIBS  = -L${MKLROOT}/lib -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lifcore -limf -lpthread -lm -ldl -liomp5
+CUDA_LDLIBS = -L/usr/local/cuda-12.9/targets/x86_64-linux/lib -lcudart
 
 ## Memory allocator switch
 ##   1 (default) : link Intel oneTBB allocator (libtbbmalloc)
@@ -24,6 +25,8 @@ ifeq ($(USE_TBBMALLOC),1)
 LDLIBS := $(TBBMALLOC_LIBS) $(LDLIBS)
 endif
 
+LDLIBS := $(CUDA_LDLIBS) $(LDLIBS)
+
 ## PGO build mode switch (ABE only; TwoPunctureABE always uses opt flags)
 ##   opt        : (default) maximum performance with PGO profile-guided optimization
 ##   instrument : PGO Phase 1 instrumentation to collect fresh profile data
@@ -44,42 +47,14 @@ INTERP_LB_FLAGS =
 endif
 
 ## Kernel implementation switch
-##   1           : use C++ rewrite of bssn_rhs and helper kernels (faster)
-##   0 (default): fall back to original Fortran kernels
-USE_CXX_KERNELS ?= 0
-
-## Z4C Cartesian RHS kernel switch
-##   1           : use C++ rewrite of Z4c_rhs (main Cartesian path faster)
-##   0 (default): use original Fortran Z4c_rhs.o
-USE_CXX_Z4C_KERNELS ?= 0
-
-## BSSN-EScalar RHS switch
-##   1           : use BSSN-EScalar C wrapper on the normal patch path
-##   0           : keep the original Fortran BSSN-EScalar RHS for precision-safe runs
-## Note: this requires USE_CXX_KERNELS=1 because the wrapper reuses the C BSSN kernel.
-USE_CXX_ESCALAR_KERNEL ?= 0
-
-## BSSN-EM RHS switch
-##   1 : use BSSN-EM C kernel (bssn_em_rhs_c.C) on the normal patch path
-##   0 : keep the original Fortran empart.f90 RHS for the EM fields (default)
-## Note: experimental, requires USE_CXX_KERNELS=1
-USE_CXX_EM_KERNEL ?= 0
-
-## Cached transfer switch
-##   auto (default): enable for BSSN vacuum, keep other paths on the safe uncached path
-##   1             : force cached Sync/Restrict/OutBd transfer on evolution hot paths
-##   0             : force the original uncached transfer path
-USE_TRANSFER_CACHE ?= auto
+##   1 (default) : use C++ rewrite of bssn_rhs and helper kernels (faster)
+##   0           : fall back to original Fortran kernels
+USE_CXX_KERNELS ?= 1
 
 ## RK4 kernel implementation switch
-##   1           : use C/C++ rewrite of rungekutta4_rout (for optimization experiments)
-##   0 (default): use original Fortran rungekutta4_rout.o
-USE_CXX_RK4 ?= 0
-
-## fmisc conservative mode switch
-##   1           : restore lower-optimization / legacy fmisc numerics
-##   0 (default): keep the optimized fmisc paths
-USE_FMISC_SAFE_MODE ?= 0
+##   1 (default) : use C/C++ rewrite of rungekutta4_rout (for optimization experiments)
+##   0           : use original Fortran rungekutta4_rout.o
+USE_CXX_RK4 ?= 1
 
 f90          = ifx
 f77          = ifx

AMSS_NCKU_source/share_func.h

@@ -46,45 +46,6 @@ static inline size_t idx_fh_F(int iF, int jF, int kF, const int ex[3]) {
     return (size_t)ii + (size_t)jj * (size_t)nx + (size_t)kk * (size_t)nx * (size_t)ny;
 }
 
-/*
- * fh 对应 Fortran: fh(0:ex1, 0:ex2, 0:ex3)
- * ord=1 => shift=0
- * iF/jF/kF 为 Fortran 索引 (0..ex)
- */
-static inline size_t idx_fh_F_ord1(int iF, int jF, int kF, const int ex[3]) {
-    const int nx = ex[0] + 1;            // ex1 + ord
-    const int ny = ex[1] + 1;
-    return (size_t)iF + (size_t)jF * (size_t)nx + (size_t)kF * (size_t)nx * (size_t)ny;
-}
-
-/*
- * fh 对应 Fortran: fh(-3:ex1, -3:ex2, -3:ex3)
- * ord=4 => shift=3
- */
-static inline size_t idx_fh_F_ord4(int iF, int jF, int kF, const int ex[3]) {
-    const int shift = 3;
-    const int nx = ex[0] + 4;            // ex1 + ord
-    const int ny = ex[1] + 4;
-    const int ii = iF + shift;           // 0..ex1+3
-    const int jj = jF + shift;           // 0..ex2+3
-    const int kk = kF + shift;           // 0..ex3+3
-    return (size_t)ii + (size_t)jj * (size_t)nx + (size_t)kk * (size_t)nx * (size_t)ny;
-}
-
-/*
- * fh 对应 Fortran: fh(-4:ex1, -4:ex2, -4:ex3)
- * ord=5 => shift=4
- */
-static inline size_t idx_fh_F_ord5(int iF, int jF, int kF, const int ex[3]) {
-    const int shift = 4;
-    const int nx = ex[0] + 5;            // ex1 + ord
-    const int ny = ex[1] + 5;
-    const int ii = iF + shift;           // 0..ex1+4
-    const int jj = jF + shift;           // 0..ex2+4
-    const int kk = kF + shift;           // 0..ex3+4
-    return (size_t)ii + (size_t)jj * (size_t)nx + (size_t)kk * (size_t)nx * (size_t)ny;
-}
-
 /*
  * func:  (1..extc1, 1..extc2, 1..extc3)   1-based in Fortran
  * funcc: (-ord+1..extc1, -ord+1..extc2, -ord+1..extc3) in Fortran
@@ -270,10 +231,7 @@ static inline void symmetry_bd(int ord,
 {
     if (ord <= 0) return;
 
-    if (ord == 1) {
-        symmetry_bd_impl(1, 0, extc, func, funcc, SoA);
-        return;
-    }
+    /* Fast paths used by current C kernels: ord=2 (derivs), ord=3 (lopsided/KO). */
     if (ord == 2) {
         symmetry_bd_impl(2, 1, extc, func, funcc, SoA);
         return;
@@ -282,91 +240,7 @@ static inline void symmetry_bd(int ord,
         symmetry_bd_impl(3, 2, extc, func, funcc, SoA);
         return;
     }
-    if (ord == 4) {
-        symmetry_bd_impl(4, 3, extc, func, funcc, SoA);
-        return;
-    }
 
     symmetry_bd_impl(ord, ord - 1, extc, func, funcc, SoA);
 }
-
-/*
- * symmetry_stbd — shell-patch (staggered boundary) ghost fill.
- *
- * Fortran: funcc(-ord+1:extc1+ord, -ord+1:extc2+ord, extc3)
- * Only 2 SoA values (x/y). No z symmetry fill.
- * Ghost on BOTH positive and negative sides of x and y.
- * Reflection uses i+2 (skips boundary) instead of i+1.
- * nx = extc1 + 2*ord, ny = extc2 + 2*ord
- */
-static inline void symmetry_stbd(int ord,
-                                 const int extc[3],
-                                 const double *func,
-                                 double *funcc,
-                                 const double SoA[2])
-{
-    const int extc1 = extc[0], extc2 = extc[1], extc3 = extc[2];
-    const int nx = extc1 + 2 * ord;
-    const int ny = extc2 + 2 * ord;
-    const int sh = ord - 1;
-    const size_t snx = (size_t)nx;
-    const size_t splane = snx * (size_t)ny;
-
-    /* 1) Copy interior: funcc(1:extc1, 1:extc2, 1:extc3) = func */
-    for (int k0 = 0; k0 < extc3; ++k0) {
-        const double *src = func + (size_t)k0 * (size_t)extc2 * (size_t)extc1;
-        const size_t kbase = (size_t)k0 * splane;
-        for (int j0 = 0; j0 < extc2; ++j0) {
-            double *dst = funcc + kbase + (size_t)(sh + j0 + 1) * snx + (size_t)(sh + 1);
-            const double *s = src + (size_t)j0 * (size_t)extc1;
-            for (int i0 = 0; i0 < extc1; ++i0) dst[i0] = s[i0];
-        }
-    }
-
-    /* 2) x-direction ghost fill */
-    const double s1 = SoA[0];
-    for (int k0 = 0; k0 < extc3; ++k0) {
-        const size_t kbase = (size_t)k0 * splane;
-        for (int j0 = 0; j0 < extc2; ++j0) {
-            const size_t off = kbase + (size_t)(sh + j0 + 1) * snx;
-            /* left side: funcc(-i) = funcc(i+2) * s1 */
-            for (int i = 0; i < ord; ++i) {
-                funcc[off + (size_t)(sh - i)]       = funcc[off + (size_t)(sh + i + 2)] * s1;
-                /* right side: funcc(extc1+1+i) = funcc(extc1-1-i) * s1 */
-                funcc[off + (size_t)(sh + extc1 + 1 + i)] = funcc[off + (size_t)(sh + extc1 - 1 - i)] * s1;
-            }
-        }
-    }
-
-    /* 3) y-direction ghost fill */
-    const double s2 = SoA[1];
-    for (int i = 0; i < nx; ++i) {
-        for (int k0 = 0; k0 < extc3; ++k0) {
-            const size_t kbase = (size_t)k0 * splane;
-            /* bottom: funcc(:,-i,:) = funcc(:,i+2,:) * s2 */
-            for (int jj = 0; jj < ord; ++jj) {
-                funcc[kbase + (size_t)(sh - jj) * snx + (size_t)i] =
-                    funcc[kbase + (size_t)(sh + jj + 2) * snx + (size_t)i] * s2;
-                /* top: funcc(:,extc2+1+jj,:) = funcc(:,extc2-1-jj,:) * s2 */
-                funcc[kbase + (size_t)(sh + extc2 + 1 + jj) * snx + (size_t)i] =
-                    funcc[kbase + (size_t)(sh + extc2 - 1 - jj) * snx + (size_t)i] * s2;
-            }
-        }
-    }
-}
-
-/*
- * Indexing for shell fh buffer: Fortran fh(-ord+1:extc1+ord, -ord+1:extc2+ord, extc3)
- * C 0-based: ii = iF + ord - 1
- * nx = extc1 + 2*ord, ny = extc2 + 2*ord
- */
-static inline size_t idx_fh_stbd(int iF, int jF, int kF, int ord, const int extc[3]) {
-    const int sh = ord - 1;
-    const int nx = extc[0] + 2 * ord;
-    const int ny = extc[1] + 2 * ord;
-    const int ii = iF + sh;
-    const int jj = jF + sh;
-    const int kk = kF - 1;  // Fortran 1-based kF → C 0-based
-    return (size_t)ii + (size_t)jj * (size_t)nx + (size_t)kk * (size_t)nx * (size_t)ny;
-}
 #endif

AMSS_NCKU_source/surface_integral.h

@@ -20,14 +20,24 @@ using namespace std;
 #include "cgh.h"
 #include "ShellPatch.h"
 #include "NullShellPatch.h"
-#include "NullShellPatch2.h"
-#include "var.h"
-#include "monitor.h"
+#include "NullShellPatch2.h"
+#include "var.h"
+#include "monitor.h"
+#include <map>
 
 class surface_integral
 {
 
 private:
+	struct SpherePointCache
+	{
+		double *pox[3];
+		SpherePointCache()
+		{
+			pox[0] = pox[1] = pox[2] = 0;
+		}
+	};
+
 	int Symmetry, factor;
 	int N_theta, N_phi; // Number of points in Theta & Phi directions
 	double dphi, dcostheta;
@@ -36,15 +46,16 @@ private:
 
 	double *nx_g, *ny_g, *nz_g; // global list of unit normals
 	int myrank, cpusize;
-	int wave_cache_spinw, wave_cache_maxl, wave_cache_modes;
-	double *wave_theta_pos, *wave_theta_neg;
-	double *wave_phi_cos, *wave_phi_sin;
-	void clear_wave_cache();
-	void build_wave_cache(int spinw, int maxl);
+	map<double, SpherePointCache> sphere_point_cache;
+	map<int, double *> shellf_cache;
+
+	void get_surface_points(double rex, double **pox);
+	double *get_shellf_buffer(int num_var);
+	void release_cached_buffers();
 
 public:
 	surface_integral(int iSymmetry);
-	~surface_integral();
+	~surface_integral();
 
 	void surf_Wave(double rex, int lev, cgh *GH, var *Rpsi4, var *Ipsi4,
 				   int spinw, int maxl, int NN, double *RP, double *IP,
@@ -82,37 +93,21 @@ public:
 								 double &, double &, double &, double &, double &, double &, double &,
 								 double &, double &, double &, double &, double &, double &,
 								 double &, double &)); // NN is the length of RP and IP
-	void surf_MassPAng(double rex, int lev, cgh *GH, var *chi, var *trK,
-					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
-					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
-					   var *Gmx, var *Gmy, var *Gmz,
-					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs,
-					   double *Rout, monitor *Monitor, bool refresh_mass_fields = true);
-	void surf_MassPAng(double rex, int lev, ShellPatch *GH, var *chi, var *trK,
-					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
-					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
-					   var *Gmx, var *Gmy, var *Gmz,
-					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs,
-					   double *Rout, monitor *Monitor, bool refresh_mass_fields = true);
-	void surf_WaveMassPAng(double rex, int lev, cgh *GH,
-					   var *Rpsi4, var *Ipsi4, int spinw, int maxl, int NN, double *RP, double *IP,
-					   var *chi, var *trK,
-					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
-					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
-					   var *Gmx, var *Gmy, var *Gmz,
-					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs,
-					   double *Rout, monitor *Monitor, bool refresh_mass_fields = true);
-	void surf_WaveMassPAng(double rex, int lev, ShellPatch *GH,
-					   var *Rpsi4, var *Ipsi4, int spinw, int maxl, int NN, double *RP, double *IP,
-					   var *chi, var *trK,
-					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
-					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
-					   var *Gmx, var *Gmy, var *Gmz,
-					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs,
-					   double *Rout, monitor *Monitor, bool refresh_mass_fields = true);
-	void surf_Wave(double rex, cgh *GH, ShellPatch *SH,
-				   var *chi, var *trK,
-				   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
+	void surf_MassPAng(double rex, int lev, cgh *GH, var *chi, var *trK,
+					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
+					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
+					   var *Gmx, var *Gmy, var *Gmz,
+					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs,
+					   double *Rout, monitor *Monitor);
+	void surf_MassPAng(double rex, int lev, ShellPatch *GH, var *chi, var *trK,
+					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
+					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
+					   var *Gmx, var *Gmy, var *Gmz,
+					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs,
+					   double *Rout, monitor *Monitor);
+	void surf_Wave(double rex, cgh *GH, ShellPatch *SH,
+				   var *chi, var *trK,
+				   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
 				   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
 				   var *chix, var *chiy, var *chiz,
 				   var *trKx, var *trKy, var *trKz,
@@ -131,12 +126,12 @@ public:
 	bool SR_Interp_Points(MyList<var> *VarList, cgh *GH, ShellPatch *SH,
 						  int NN, double **XX, double *Shellf);
 
-	void surf_MassPAng(double rex, int lev, cgh *GH, var *chi, var *trK,
-					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
-					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
-					   var *Gmx, var *Gmy, var *Gmz,
-					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs, // temparay memory for mass^i
-					   double *Rout, monitor *Monitor, MPI_Comm Comm_here, bool refresh_mass_fields = true);
+	void surf_MassPAng(double rex, int lev, cgh *GH, var *chi, var *trK,
+					   var *gxx, var *gxy, var *gxz, var *gyy, var *gyz, var *gzz,
+					   var *Axx, var *Axy, var *Axz, var *Ayy, var *Ayz, var *Azz,
+					   var *Gmx, var *Gmy, var *Gmz,
+					   var *Sfx_rhs, var *Sfy_rhs, var *Sfz_rhs, // temparay memory for mass^i
+					   double *Rout, monitor *Monitor, MPI_Comm Comm_here);
 	void surf_Wave(double rex, int lev, cgh *GH, var *Rpsi4, var *Ipsi4,
 				   int spinw, int maxl, int NN, double *RP, double *IP,
 				   monitor *Monitor, MPI_Comm Comm_here);

AMSS_NCKU_source/z4c_rhs_c.C

@@ -1,901 +0,0 @@
-#include "macrodef.h"
-#include "bssn_rhs.h"
-#include "fmisc.h"
-#include "ricci_gamma.h"
-#include "share_func.h"
-#include "tool.h"
-#include <vector>
-
-#ifdef fortran1
-#define f_constraint_bssn constraint_bssn
-#define f_z4c_rhs_point z4c_rhs_point
-#endif
-#ifdef fortran2
-#define f_constraint_bssn CONSTRAINT_BSSN
-#define f_z4c_rhs_point Z4C_RHS_POINT
-#endif
-#ifdef fortran3
-#define f_constraint_bssn constraint_bssn_
-#define f_z4c_rhs_point z4c_rhs_point_
-#endif
-
-extern "C" void f_constraint_bssn(int *, double *, double *, double *,
-                                  double *, double *,
-                                  double *, double *, double *, double *, double *, double *,
-                                  double *, double *, double *, double *, double *, double *,
-                                  double *, double *, double *,
-                                  double *, double *, double *, double *, double *, double *, double *, double *, double *, double *,
-                                  double *, double *, double *, double *, double *, double *,
-                                  double *, double *, double *, double *, double *, double *,
-                                  double *, double *, double *, double *, double *, double *,
-                                  double *, double *, double *, double *, double *, double *, double *, double *,
-                                  double *, double *, double *,
-                                  int &);
-
-extern "C" void f_z4c_rhs_point(
-    double &A11,
-    double &A12,
-    double &A13,
-    double &A22,
-    double &A23,
-    double &A33,
-    double &alpha,
-    double &B1,
-    double &B2,
-    double &B3,
-    double &beta1,
-    double &beta2,
-    double &beta3,
-    double &chi,
-    double &chiDivFloor,
-    double &da1,
-    double &dA111,
-    double &dA112,
-    double &dA113,
-    double &dA122,
-    double &dA123,
-    double &dA133,
-    double &da2,
-    double &dA211,
-    double &dA212,
-    double &dA213,
-    double &dA222,
-    double &dA223,
-    double &dA233,
-    double &da3,
-    double &dA311,
-    double &dA312,
-    double &dA313,
-    double &dA322,
-    double &dA323,
-    double &dA333,
-    double &db11,
-    double &dB11,
-    double &db12,
-    double &dB12,
-    double &db13,
-    double &dB13,
-    double &db21,
-    double &dB21,
-    double &db22,
-    double &dB22,
-    double &db23,
-    double &dB23,
-    double &db31,
-    double &dB31,
-    double &db32,
-    double &dB32,
-    double &db33,
-    double &dB33,
-    double &dchi1,
-    double &dchi2,
-    double &dchi3,
-    double &dda11,
-    double &dda12,
-    double &dda13,
-    double &dda22,
-    double &dda23,
-    double &dda33,
-    double &ddb111,
-    double &ddb112,
-    double &ddb113,
-    double &ddb121,
-    double &ddb122,
-    double &ddb123,
-    double &ddb131,
-    double &ddb132,
-    double &ddb133,
-    double &ddb221,
-    double &ddb222,
-    double &ddb223,
-    double &ddb231,
-    double &ddb232,
-    double &ddb233,
-    double &ddb331,
-    double &ddb332,
-    double &ddb333,
-    double &ddchi11,
-    double &ddchi12,
-    double &ddchi13,
-    double &ddchi22,
-    double &ddchi23,
-    double &ddchi33,
-    double &deldelg1111,
-    double &deldelg1112,
-    double &deldelg1113,
-    double &deldelg1122,
-    double &deldelg1123,
-    double &deldelg1133,
-    double &deldelg1211,
-    double &deldelg1212,
-    double &deldelg1213,
-    double &deldelg1222,
-    double &deldelg1223,
-    double &deldelg1233,
-    double &deldelg1311,
-    double &deldelg1312,
-    double &deldelg1313,
-    double &deldelg1322,
-    double &deldelg1323,
-    double &deldelg1333,
-    double &deldelg2211,
-    double &deldelg2212,
-    double &deldelg2213,
-    double &deldelg2222,
-    double &deldelg2223,
-    double &deldelg2233,
-    double &deldelg2311,
-    double &deldelg2312,
-    double &deldelg2313,
-    double &deldelg2322,
-    double &deldelg2323,
-    double &deldelg2333,
-    double &deldelg3311,
-    double &deldelg3312,
-    double &deldelg3313,
-    double &deldelg3322,
-    double &deldelg3323,
-    double &deldelg3333,
-    double &delG11,
-    double &delg111,
-    double &delg112,
-    double &delg113,
-    double &delG12,
-    double &delg122,
-    double &delg123,
-    double &delG13,
-    double &delg133,
-    double &delG21,
-    double &delg211,
-    double &delg212,
-    double &delg213,
-    double &delG22,
-    double &delg222,
-    double &delg223,
-    double &delG23,
-    double &delg233,
-    double &delG31,
-    double &delg311,
-    double &delg312,
-    double &delg313,
-    double &delG32,
-    double &delg322,
-    double &delg323,
-    double &delG33,
-    double &delg333,
-    double &dKhat1,
-    double &dKhat2,
-    double &dKhat3,
-    double &dTheta1,
-    double &dTheta2,
-    double &dTheta3,
-    double &G1,
-    double &g11,
-    double &g12,
-    double &g13,
-    double &G2,
-    double &g22,
-    double &g23,
-    double &G3,
-    double &g33,
-    double &kappa1,
-    double &kappa2,
-    double &Khat,
-    double &rA11,
-    double &rA12,
-    double &rA13,
-    double &rA22,
-    double &rA23,
-    double &rA33,
-    double &rchi,
-    double &rG1,
-    double &rg11,
-    double &rg12,
-    double &rg13,
-    double &rG2,
-    double &rg22,
-    double &rg23,
-    double &rG3,
-    double &rg33,
-    double &rKhat,
-    double &rTheta,
-    double &Theta);
-
-static inline void z4c_contract_gamma(
-    const double gxx, const double gxy, const double gxz,
-    const double gyy, const double gyz, const double gzz,
-    const double gxxx, const double gxyx, const double gxzx,
-    const double gyyx, const double gyzx, const double gzzx,
-    const double gxxy, const double gxyy, const double gxzy,
-    const double gyyy, const double gyzy, const double gzzy,
-    const double gxxz, const double gxyz, const double gxzz,
-    const double gyyz, const double gyzz, const double gzzz,
-    double &Gamxa, double &Gamya, double &Gamza)
-{
-    double det = gxx * gyy * gzz + gxy * gyz * gxz + gxz * gxy * gyz -
-                 gxz * gyy * gxz - gxy * gxy * gzz - gxx * gyz * gyz;
-    const double gupxx = (gyy * gzz - gyz * gyz) / det;
-    const double gupxy = -(gxy * gzz - gyz * gxz) / det;
-    const double gupxz = (gxy * gyz - gyy * gxz) / det;
-    const double gupyy = (gxx * gzz - gxz * gxz) / det;
-    const double gupyz = -(gxx * gyz - gxy * gxz) / det;
-    const double gupzz = (gxx * gyy - gxy * gxy) / det;
-
-    const double Gamxxx = 0.5 * (gupxx * gxxx + gupxy * (2.0 * gxyx - gxxy) + gupxz * (2.0 * gxzx - gxxz));
-    const double Gamyxx = 0.5 * (gupxy * gxxx + gupyy * (2.0 * gxyx - gxxy) + gupyz * (2.0 * gxzx - gxxz));
-    const double Gamzxx = 0.5 * (gupxz * gxxx + gupyz * (2.0 * gxyx - gxxy) + gupzz * (2.0 * gxzx - gxxz));
-
-    const double Gamxyy = 0.5 * (gupxx * (2.0 * gxyy - gyyx) + gupxy * gyyy + gupxz * (2.0 * gyzy - gyyz));
-    const double Gamyyy = 0.5 * (gupxy * (2.0 * gxyy - gyyx) + gupyy * gyyy + gupyz * (2.0 * gyzy - gyyz));
-    const double Gamzyy = 0.5 * (gupxz * (2.0 * gxyy - gyyx) + gupyz * gyyy + gupzz * (2.0 * gyzy - gyyz));
-
-    const double Gamxzz = 0.5 * (gupxx * (2.0 * gxzz - gzzx) + gupxy * (2.0 * gyzz - gzzy) + gupxz * gzzz);
-    const double Gamyzz = 0.5 * (gupxy * (2.0 * gxzz - gzzx) + gupyy * (2.0 * gyzz - gzzy) + gupyz * gzzz);
-    const double Gamzzz = 0.5 * (gupxz * (2.0 * gxzz - gzzx) + gupyz * (2.0 * gyzz - gzzy) + gupzz * gzzz);
-
-    const double Gamxxy = 0.5 * (gupxx * gxxy + gupxy * gyyx + gupxz * (gxzy + gyzx - gxyz));
-    const double Gamyxy = 0.5 * (gupxy * gxxy + gupyy * gyyx + gupyz * (gxzy + gyzx - gxyz));
-    const double Gamzxy = 0.5 * (gupxz * gxxy + gupyz * gyyx + gupzz * (gxzy + gyzx - gxyz));
-
-    const double Gamxxz = 0.5 * (gupxx * gxxz + gupxy * (gxyz + gyzx - gxzy) + gupxz * gzzx);
-    const double Gamyxz = 0.5 * (gupxy * gxxz + gupyy * (gxyz + gyzx - gxzy) + gupyz * gzzx);
-    const double Gamzxz = 0.5 * (gupxz * gxxz + gupyz * (gxyz + gyzx - gxzy) + gupzz * gzzx);
-
-    const double Gamxyz = 0.5 * (gupxx * (gxyz + gxzy - gyzx) + gupxy * gyyz + gupxz * gzzy);
-    const double Gamyyz = 0.5 * (gupxy * (gxyz + gxzy - gyzx) + gupyy * gyyz + gupyz * gzzy);
-    const double Gamzyz = 0.5 * (gupxz * (gxyz + gxzy - gyzx) + gupyz * gyyz + gupzz * gzzy);
-
-    Gamxa = gupxx * Gamxxx + gupyy * Gamxyy + gupzz * Gamxzz +
-            2.0 * (gupxy * Gamxxy + gupxz * Gamxxz + gupyz * Gamxyz);
-    Gamya = gupxx * Gamyxx + gupyy * Gamyyy + gupzz * Gamyzz +
-            2.0 * (gupxy * Gamyxy + gupxz * Gamyxz + gupyz * Gamyyz);
-    Gamza = gupxx * Gamzxx + gupyy * Gamzyy + gupzz * Gamzzz +
-            2.0 * (gupxy * Gamzxy + gupxz * Gamzxz + gupyz * Gamzyz);
-}
-
-static int compute_rhs_z4c_cartesian(
-    int *ex, double &T, double *X, double *Y, double *Z,
-    double *chi_state, double *chi_constraints, double *trK,
-    double *dxx, double *gxy, double *gxz, double *dyy, double *gyz, double *dzz,
-    double *Axx, double *Axy, double *Axz, double *Ayy, double *Ayz, double *Azz,
-    double *Gamx, double *Gamy, double *Gamz,
-    double *Lap, double *betax, double *betay, double *betaz,
-    double *dtSfx, double *dtSfy, double *dtSfz,
-    double *TZ,
-    double *chi_rhs, double *trK_rhs,
-    double *gxx_rhs, double *gxy_rhs, double *gxz_rhs, double *gyy_rhs, double *gyz_rhs, double *gzz_rhs,
-    double *Axx_rhs, double *Axy_rhs, double *Axz_rhs, double *Ayy_rhs, double *Ayz_rhs, double *Azz_rhs,
-    double *Gamx_rhs, double *Gamy_rhs, double *Gamz_rhs,
-    double *Lap_rhs, double *betax_rhs, double *betay_rhs, double *betaz_rhs,
-    double *dtSfx_rhs, double *dtSfy_rhs, double *dtSfz_rhs,
-    double *TZ_rhs,
-    double *rho, double *Sx, double *Sy, double *Sz,
-    double *Sxx, double *Sxy, double *Sxz, double *Syy, double *Syz, double *Szz,
-    double *Gamxxx, double *Gamxxy, double *Gamxxz, double *Gamxyy, double *Gamxyz, double *Gamxzz,
-    double *Gamyxx, double *Gamyxy, double *Gamyxz, double *Gamyyy, double *Gamyyz, double *Gamyzz,
-    double *Gamzxx, double *Gamzxy, double *Gamzxz, double *Gamzyy, double *Gamzyz, double *Gamzzz,
-    double *Rxx, double *Rxy, double *Rxz, double *Ryy, double *Ryz, double *Rzz,
-    double *Hcon, double *Mxcon, double *Mycon, double *Mzcon, double *Gmxcon, double *Gmycon, double *Gmzcon,
-    int &Symmetry, int &Lev, double &eps, int &co)
-{
-    (void)T;
-
-    const int nx = ex[0];
-    const int ny = ex[1];
-    const int nz = ex[2];
-    const int all = nx * ny * nz;
-
-    double alpn1[all], chin1[all], gxx[all], gyy[all], gzz[all];
-    double chix[all], chiy[all], chiz[all], chixx[all], chixy[all], chixz[all], chiyy[all], chiyz[all], chizz[all];
-    double gxxx[all], gxyx[all], gxzx[all], gyyx[all], gyzx[all], gzzx[all];
-    double gxxy[all], gxyy[all], gxzy[all], gyyy[all], gyzy[all], gzzy[all];
-    double gxxz[all], gxyz[all], gxzz[all], gyyz[all], gyzz[all], gzzz[all];
-    double gxxxx[all], gxxxy[all], gxxxz[all], gxxyy[all], gxxyz[all], gxxzz[all];
-    double gxyxx[all], gxyxy[all], gxyxz[all], gxyyy[all], gxyyz[all], gxyzz[all];
-    double gxzxx[all], gxzxy[all], gxzxz[all], gxzyy[all], gxzyz[all], gxzzz[all];
-    double gyyxx[all], gyyxy[all], gyyxz[all], gyyyy[all], gyyyz[all], gyyzz[all];
-    double gyzxx[all], gyzxy[all], gyzxz[all], gyzyy[all], gyzyz[all], gyzzz[all];
-    double gzzxx[all], gzzxy[all], gzzxz[all], gzzyy[all], gzzyz[all], gzzzz[all];
-    double Lapx[all], Lapy[all], Lapz[all], Lapxx[all], Lapxy[all], Lapxz[all], Lapyy[all], Lapyz[all], Lapzz[all];
-    double betaxx[all], betaxy[all], betaxz[all], betayx[all], betayy[all], betayz[all], betazx[all], betazy[all], betazz[all];
-    double dBxx[all], dBxy[all], dBxz[all], dByx[all], dByy[all], dByz[all], dBzx[all], dBzy[all], dBzz[all];
-    double sfxxx[all], sfxxy[all], sfxxz[all], sfxyy[all], sfxyz[all], sfxzz[all];
-    double sfyxx[all], sfyxy[all], sfyxz[all], sfyyy[all], sfyyz[all], sfyzz[all];
-    double sfzxx[all], sfzxy[all], sfzxz[all], sfzyy[all], sfzyz[all], sfzzz[all];
-    double Gamxx[all], Gamxy[all], Gamxz[all], Gamyx[all], Gamyy[all], Gamyz[all], Gamzx[all], Gamzy[all], Gamzz[all];
-    double Kx[all], Ky[all], Kz[all], TZx[all], TZy[all], TZz[all];
-    double Axxx[all], Axxy[all], Axxz[all], Axyx[all], Axyy[all], Axyz[all];
-    double Axzx[all], Axzy[all], Axzz[all], Ayyx[all], Ayyy[all], Ayyz[all];
-    double Ayzx[all], Ayzy[all], Ayzz[all], Azzx[all], Azzy[all], Azzz[all];
-#if (GAUGE == 2 || GAUGE == 3 || GAUGE == 4 || GAUGE == 5)
-    double reta[all];
-#endif
-
-    const double SSS[3] = {1.0, 1.0, 1.0};
-    const double AAS[3] = {-1.0, -1.0, 1.0};
-    const double ASA[3] = {-1.0, 1.0, -1.0};
-    const double SAA[3] = {1.0, -1.0, -1.0};
-    const double ASS[3] = {-1.0, 1.0, 1.0};
-    const double SAS[3] = {1.0, -1.0, 1.0};
-    const double SSA[3] = {1.0, 1.0, -1.0};
-
-    const double ONE = 1.0;
-    const double TWO = 2.0;
-    const double ZEO = 0.0;
-    double chiDivfloor = 1.0e-5;
-
-    double kappa1 = 2.0e-2;
-    double kappa2 = 0.0;
-    double FF = 0.75;
-    double eta = 2.0;
-
-    for (int idx = 0; idx < all; ++idx)
-    {
-        alpn1[idx] = Lap[idx] + ONE;
-        chin1[idx] = chi_state[idx] + ONE;
-        gxx[idx] = dxx[idx] + ONE;
-        gyy[idx] = dyy[idx] + ONE;
-        gzz[idx] = dzz[idx] + ONE;
-    }
-
-    fderivs(ex, betax, betaxx, betaxy, betaxz, X, Y, Z, -1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, betay, betayx, betayy, betayz, X, Y, Z, 1.0, -1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, betaz, betazx, betazy, betazz, X, Y, Z, 1.0, 1.0, -1.0, Symmetry, Lev);
-    fderivs(ex, dtSfx, dBxx, dBxy, dBxz, X, Y, Z, -1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, dtSfy, dByx, dByy, dByz, X, Y, Z, 1.0, -1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, dtSfz, dBzx, dBzy, dBzz, X, Y, Z, 1.0, 1.0, -1.0, Symmetry, Lev);
-    fderivs(ex, chi_state, chix, chiy, chiz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, dxx, gxxx, gxxy, gxxz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, gxy, gxyx, gxyy, gxyz, X, Y, Z, -1.0, -1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, gxz, gxzx, gxzy, gxzz, X, Y, Z, -1.0, 1.0, -1.0, Symmetry, Lev);
-    fderivs(ex, dyy, gyyx, gyyy, gyyz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, gyz, gyzx, gyzy, gyzz, X, Y, Z, 1.0, -1.0, -1.0, Symmetry, Lev);
-    fderivs(ex, dzz, gzzx, gzzy, gzzz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-
-    fdderivs(ex, dxx, gxxxx, gxxxy, gxxxz, gxxyy, gxxyz, gxxzz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, dyy, gyyxx, gyyxy, gyyxz, gyyyy, gyyyz, gyyzz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, dzz, gzzxx, gzzxy, gzzxz, gzzyy, gzzyz, gzzzz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, gxy, gxyxx, gxyxy, gxyxz, gxyyy, gxyyz, gxyzz, X, Y, Z, -1.0, -1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, gxz, gxzxx, gxzxy, gxzxz, gxzyy, gxzyz, gxzzz, X, Y, Z, -1.0, 1.0, -1.0, Symmetry, Lev);
-    fdderivs(ex, gyz, gyzxx, gyzxy, gyzxz, gyzyy, gyzyz, gyzzz, X, Y, Z, 1.0, -1.0, -1.0, Symmetry, Lev);
-
-    fderivs(ex, Gamx, Gamxx, Gamxy, Gamxz, X, Y, Z, -1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, Gamy, Gamyx, Gamyy, Gamyz, X, Y, Z, 1.0, -1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, Gamz, Gamzx, Gamzy, Gamzz, X, Y, Z, 1.0, 1.0, -1.0, Symmetry, Lev);
-
-    fderivs(ex, Lap, Lapx, Lapy, Lapz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, trK, Kx, Ky, Kz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, TZ, TZx, TZy, TZz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-
-    fdderivs(ex, betax, sfxxx, sfxxy, sfxxz, sfxyy, sfxyz, sfxzz, X, Y, Z, -1.0, 1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, betay, sfyxx, sfyxy, sfyxz, sfyyy, sfyyz, sfyzz, X, Y, Z, 1.0, -1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, betaz, sfzxx, sfzxy, sfzxz, sfzyy, sfzyz, sfzzz, X, Y, Z, 1.0, 1.0, -1.0, Symmetry, Lev);
-
-    fdderivs(ex, chi_state, chixx, chixy, chixz, chiyy, chiyz, chizz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fdderivs(ex, Lap, Lapxx, Lapxy, Lapxz, Lapyy, Lapyz, Lapzz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-
-    fderivs(ex, Axx, Axxx, Axxy, Axxz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, Axy, Axyx, Axyy, Axyz, X, Y, Z, -1.0, -1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, Axz, Axzx, Axzy, Axzz, X, Y, Z, -1.0, 1.0, -1.0, Symmetry, Lev);
-    fderivs(ex, Ayy, Ayyx, Ayyy, Ayyz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-    fderivs(ex, Ayz, Ayzx, Ayzy, Ayzz, X, Y, Z, 1.0, -1.0, -1.0, Symmetry, Lev);
-    fderivs(ex, Azz, Azzx, Azzy, Azzz, X, Y, Z, 1.0, 1.0, 1.0, Symmetry, Lev);
-
-    for (int idx = 0; idx < all; ++idx)
-    {
-        double point_kappa1 = 0.0;
-        f_z4c_rhs_point(
-            Axx[idx], Axy[idx], Axz[idx], Ayy[idx], Ayz[idx], Azz[idx],
-            alpn1[idx], dtSfx[idx], dtSfy[idx], dtSfz[idx],
-            betax[idx], betay[idx], betaz[idx],
-            chin1[idx], chiDivfloor,
-            Lapx[idx],
-            Axxx[idx], Axyx[idx], Axzx[idx], Ayyx[idx], Ayzx[idx], Azzx[idx],
-            Lapy[idx],
-            Axxy[idx], Axyy[idx], Axzy[idx], Ayyy[idx], Ayzy[idx], Azzy[idx],
-            Lapz[idx],
-            Axxz[idx], Axyz[idx], Axzz[idx], Ayyz[idx], Ayzz[idx], Azzz[idx],
-            betaxx[idx], dBxx[idx], betayx[idx], dByx[idx], betazx[idx], dBzx[idx],
-            betaxy[idx], dBxy[idx], betayy[idx], dByy[idx], betazy[idx], dBzy[idx],
-            betaxz[idx], dBxz[idx], betayz[idx], dByz[idx], betazz[idx], dBzz[idx],
-            chix[idx], chiy[idx], chiz[idx],
-            Lapxx[idx], Lapxy[idx], Lapxz[idx], Lapyy[idx], Lapyz[idx], Lapzz[idx],
-            sfxxx[idx], sfyxx[idx], sfzxx[idx],
-            sfxxy[idx], sfyxy[idx], sfzxy[idx],
-            sfxxz[idx], sfyxz[idx], sfzxz[idx],
-            sfxyy[idx], sfyyy[idx], sfzyy[idx],
-            sfxyz[idx], sfyyz[idx], sfzyz[idx],
-            sfxzz[idx], sfyzz[idx], sfzzz[idx],
-            chixx[idx], chixy[idx], chixz[idx], chiyy[idx], chiyz[idx], chizz[idx],
-            gxxxx[idx], gxyxx[idx], gxzxx[idx], gyyxx[idx], gyzxx[idx], gzzxx[idx],
-            gxxxy[idx], gxyxy[idx], gxzxy[idx], gyyxy[idx], gyzxy[idx], gzzxy[idx],
-            gxxxz[idx], gxyxz[idx], gxzxz[idx], gyyxz[idx], gyzxz[idx], gzzxz[idx],
-            gxxyy[idx], gxyyy[idx], gxzyy[idx], gyyyy[idx], gyzyy[idx], gzzyy[idx],
-            gxxyz[idx], gxyyz[idx], gxzyz[idx], gyyyz[idx], gyzyz[idx], gzzyz[idx],
-            gxxzz[idx], gxyzz[idx], gxzzz[idx], gyyzz[idx], gyzzz[idx], gzzzz[idx],
-            Gamxx[idx], gxxx[idx], gxyx[idx], gxzx[idx],
-            Gamyx[idx], gyyx[idx], gyzx[idx],
-            Gamzx[idx], gzzx[idx],
-            Gamxy[idx], gxxy[idx], gxyy[idx], gxzy[idx],
-            Gamyy[idx], gyyy[idx], gyzy[idx],
-            Gamzy[idx], gzzy[idx],
-            Gamxz[idx], gxxz[idx], gxyz[idx], gxzz[idx],
-            Gamyz[idx], gyyz[idx], gyzz[idx],
-            Gamzz[idx], gzzz[idx],
-            Kx[idx], Ky[idx], Kz[idx],
-            TZx[idx], TZy[idx], TZz[idx],
-            Gamx[idx], gxx[idx], gxy[idx], gxz[idx],
-            Gamy[idx], gyy[idx], gyz[idx],
-            Gamz[idx], gzz[idx],
-            point_kappa1, kappa2,
-            trK[idx],
-            Axx_rhs[idx], Axy_rhs[idx], Axz_rhs[idx], Ayy_rhs[idx], Ayz_rhs[idx], Azz_rhs[idx],
-            chi_rhs[idx],
-            Gamx_rhs[idx], gxx_rhs[idx], gxy_rhs[idx], gxz_rhs[idx],
-            Gamy_rhs[idx], gyy_rhs[idx], gyz_rhs[idx],
-            Gamz_rhs[idx], gzz_rhs[idx], trK_rhs[idx], TZ_rhs[idx], TZ[idx]);
-    }
-
-    for (int idx = 0; idx < all; ++idx)
-        Lap_rhs[idx] = -TWO * alpn1[idx] * trK[idx];
-
-#if (GAUGE == 0)
-    for (int idx = 0; idx < all; ++idx)
-    {
-        betax_rhs[idx] = FF * dtSfx[idx];
-        betay_rhs[idx] = FF * dtSfy[idx];
-        betaz_rhs[idx] = FF * dtSfz[idx];
-        dtSfx_rhs[idx] = Gamx_rhs[idx] - eta * dtSfx[idx];
-        dtSfy_rhs[idx] = Gamy_rhs[idx] - eta * dtSfy[idx];
-        dtSfz_rhs[idx] = Gamz_rhs[idx] - eta * dtSfz[idx];
-    }
-#elif (GAUGE == 1)
-    for (int idx = 0; idx < all; ++idx)
-    {
-        betax_rhs[idx] = Gamx[idx] - eta * betax[idx];
-        betay_rhs[idx] = Gamy[idx] - eta * betay[idx];
-        betaz_rhs[idx] = Gamz[idx] - eta * betaz[idx];
-        dtSfx_rhs[idx] = ZEO;
-        dtSfy_rhs[idx] = ZEO;
-        dtSfz_rhs[idx] = ZEO;
-    }
-#elif (GAUGE == 2)
-    /* Variable-eta gamma-driver, chi-sqrt denominator */
-    for (int idx = 0; idx < all; ++idx)
-    {
-        const double chin1i = chin1[idx];
-        const double det = gxx[idx] * gyy[idx] * gzz[idx]
-                         + gxy[idx] * gyz[idx] * gxz[idx] * 2.0
-                         - gxz[idx] * gyy[idx] * gxz[idx]
-                         - gxy[idx] * gxy[idx] * gzz[idx]
-                         - gxx[idx] * gyz[idx] * gyz[idx];
-        const double idet = ONE / det;
-        const double upxx = (gyy[idx] * gzz[idx] - gyz[idx] * gyz[idx]) * idet;
-        const double upxy = -(gxy[idx] * gzz[idx] - gyz[idx] * gxz[idx]) * idet;
-        const double upxz = (gxy[idx] * gyz[idx] - gyy[idx] * gxz[idx]) * idet;
-        const double upyy = (gxx[idx] * gzz[idx] - gxz[idx] * gxz[idx]) * idet;
-        const double upyz = -(gxx[idx] * gyz[idx] - gxy[idx] * gxz[idx]) * idet;
-        const double upzz = (gxx[idx] * gyy[idx] - gxy[idx] * gxy[idx]) * idet;
-        const double grdchi2 =
-            upxx * chix[idx] * chix[idx] + upyy * chiy[idx] * chiy[idx] + upzz * chiz[idx] * chiz[idx]
-          + TWO * (upxy * chix[idx] * chiy[idx] + upxz * chix[idx] * chiz[idx] + upyz * chiy[idx] * chiz[idx]);
-        const double sqchi = sqrt(chin1i);
-        reta[idx] = 1.31 / TWO * sqrt(grdchi2 / chin1i) / ((ONE - sqchi) * (ONE - sqchi));
-        betax_rhs[idx] = FF * dtSfx[idx];
-        betay_rhs[idx] = FF * dtSfy[idx];
-        betaz_rhs[idx] = FF * dtSfz[idx];
-        dtSfx_rhs[idx] = Gamx_rhs[idx] - reta[idx] * dtSfx[idx];
-        dtSfy_rhs[idx] = Gamy_rhs[idx] - reta[idx] * dtSfy[idx];
-        dtSfz_rhs[idx] = Gamz_rhs[idx] - reta[idx] * dtSfz[idx];
-    }
-#elif (GAUGE == 3)
-    /* Variable-eta gamma-driver, chi-linear denominator */
-    for (int idx = 0; idx < all; ++idx)
-    {
-        const double chin1i = chin1[idx];
-        const double det = gxx[idx] * gyy[idx] * gzz[idx]
-                         + gxy[idx] * gyz[idx] * gxz[idx] * 2.0
-                         - gxz[idx] * gyy[idx] * gxz[idx]
-                         - gxy[idx] * gxy[idx] * gzz[idx]
-                         - gxx[idx] * gyz[idx] * gyz[idx];
-        const double idet = ONE / det;
-        const double upxx = (gyy[idx] * gzz[idx] - gyz[idx] * gyz[idx]) * idet;
-        const double upxy = -(gxy[idx] * gzz[idx] - gyz[idx] * gxz[idx]) * idet;
-        const double upxz = (gxy[idx] * gyz[idx] - gyy[idx] * gxz[idx]) * idet;
-        const double upyy = (gxx[idx] * gzz[idx] - gxz[idx] * gxz[idx]) * idet;
-        const double upyz = -(gxx[idx] * gyz[idx] - gxy[idx] * gxz[idx]) * idet;
-        const double upzz = (gxx[idx] * gyy[idx] - gxy[idx] * gxy[idx]) * idet;
-        const double grdchi2 =
-            upxx * chix[idx] * chix[idx] + upyy * chiy[idx] * chiy[idx] + upzz * chiz[idx] * chiz[idx]
-          + TWO * (upxy * chix[idx] * chiy[idx] + upxz * chix[idx] * chiz[idx] + upyz * chiy[idx] * chiz[idx]);
-        reta[idx] = 1.31 / TWO * sqrt(grdchi2 / chin1i) / ((ONE - chin1i) * (ONE - chin1i));
-        betax_rhs[idx] = FF * dtSfx[idx];
-        betay_rhs[idx] = FF * dtSfy[idx];
-        betaz_rhs[idx] = FF * dtSfz[idx];
-        dtSfx_rhs[idx] = Gamx_rhs[idx] - reta[idx] * dtSfx[idx];
-        dtSfy_rhs[idx] = Gamy_rhs[idx] - reta[idx] * dtSfy[idx];
-        dtSfz_rhs[idx] = Gamz_rhs[idx] - reta[idx] * dtSfz[idx];
-    }
-#elif (GAUGE == 4)
-    /* Variable-eta gamma-driver, first-order, chi-sqrt denominator */
-    for (int idx = 0; idx < all; ++idx)
-    {
-        const double chin1i = chin1[idx];
-        const double det = gxx[idx] * gyy[idx] * gzz[idx]
-                         + gxy[idx] * gyz[idx] * gxz[idx] * 2.0
-                         - gxz[idx] * gyy[idx] * gxz[idx]
-                         - gxy[idx] * gxy[idx] * gzz[idx]
-                         - gxx[idx] * gyz[idx] * gyz[idx];
-        const double idet = ONE / det;
-        const double upxx = (gyy[idx] * gzz[idx] - gyz[idx] * gyz[idx]) * idet;
-        const double upxy = -(gxy[idx] * gzz[idx] - gyz[idx] * gxz[idx]) * idet;
-        const double upxz = (gxy[idx] * gyz[idx] - gyy[idx] * gxz[idx]) * idet;
-        const double upyy = (gxx[idx] * gzz[idx] - gxz[idx] * gxz[idx]) * idet;
-        const double upyz = -(gxx[idx] * gyz[idx] - gxy[idx] * gxz[idx]) * idet;
-        const double upzz = (gxx[idx] * gyy[idx] - gxy[idx] * gxy[idx]) * idet;
-        const double grdchi2 =
-            upxx * chix[idx] * chix[idx] + upyy * chiy[idx] * chiy[idx] + upzz * chiz[idx] * chiz[idx]
-          + TWO * (upxy * chix[idx] * chiy[idx] + upxz * chix[idx] * chiz[idx] + upyz * chiy[idx] * chiz[idx]);
-        const double sqchi = sqrt(chin1i);
-        reta[idx] = 1.31 / TWO * sqrt(grdchi2 / chin1i) / ((ONE - sqchi) * (ONE - sqchi));
-        betax_rhs[idx] = Gamx_rhs[idx] - reta[idx] * betax[idx];
-        betay_rhs[idx] = Gamy_rhs[idx] - reta[idx] * betay[idx];
-        betaz_rhs[idx] = Gamz_rhs[idx] - reta[idx] * betaz[idx];
-        dtSfx_rhs[idx] = ZEO;
-        dtSfy_rhs[idx] = ZEO;
-        dtSfz_rhs[idx] = ZEO;
-    }
-#elif (GAUGE == 5)
-    /* Variable-eta gamma-driver, first-order, chi-linear denominator */
-    for (int idx = 0; idx < all; ++idx)
-    {
-        const double chin1i = chin1[idx];
-        const double det = gxx[idx] * gyy[idx] * gzz[idx]
-                         + gxy[idx] * gyz[idx] * gxz[idx] * 2.0
-                         - gxz[idx] * gyy[idx] * gxz[idx]
-                         - gxy[idx] * gxy[idx] * gzz[idx]
-                         - gxx[idx] * gyz[idx] * gyz[idx];
-        const double idet = ONE / det;
-        const double upxx = (gyy[idx] * gzz[idx] - gyz[idx] * gyz[idx]) * idet;
-        const double upxy = -(gxy[idx] * gzz[idx] - gyz[idx] * gxz[idx]) * idet;
-        const double upxz = (gxy[idx] * gyz[idx] - gyy[idx] * gxz[idx]) * idet;
-        const double upyy = (gxx[idx] * gzz[idx] - gxz[idx] * gxz[idx]) * idet;
-        const double upyz = -(gxx[idx] * gyz[idx] - gxy[idx] * gxz[idx]) * idet;
-        const double upzz = (gxx[idx] * gyy[idx] - gxy[idx] * gxy[idx]) * idet;
-        const double grdchi2 =
-            upxx * chix[idx] * chix[idx] + upyy * chiy[idx] * chiy[idx] + upzz * chiz[idx] * chiz[idx]
-          + TWO * (upxy * chix[idx] * chiy[idx] + upxz * chix[idx] * chiz[idx] + upyz * chiy[idx] * chiz[idx]);
-        reta[idx] = 1.31 / TWO * sqrt(grdchi2 / chin1i) / ((ONE - chin1i) * (ONE - chin1i));
-        betax_rhs[idx] = Gamx_rhs[idx] - reta[idx] * betax[idx];
-        betay_rhs[idx] = Gamy_rhs[idx] - reta[idx] * betay[idx];
-        betaz_rhs[idx] = Gamz_rhs[idx] - reta[idx] * betaz[idx];
-        dtSfx_rhs[idx] = ZEO;
-        dtSfy_rhs[idx] = ZEO;
-        dtSfz_rhs[idx] = ZEO;
-    }
-#elif (GAUGE == 6 || GAUGE == 7)
-    {
-        /* Jason's position-dependent damping: rational (6) or exponential (7) */
-        int BHN = 0;
-        double Porg[9] = {0.0};
-        double Mass[3] = {0.0};
-    #ifdef fortran1
-        extern "C" { void getpbh(int &, double *, double *); }
-    #elif defined(fortran2)
-        extern "C" { void GETPBH(int &, double *, double *); }
-    #else
-        extern "C" { void getpbh_(int &, double *, double *); }
-    #endif
-        {
-    #ifdef fortran1
-            getpbh(BHN, Porg, Mass);
-    #elif defined(fortran2)
-            GETPBH(BHN, Porg, Mass);
-    #else
-            getpbh_(BHN, Porg, Mass);
-    #endif
-        }
-        if (BHN == 2)
-        {
-            const double M = Mass[0] + Mass[1];
-            const double A = 2.0 / M;
-            const double w1 = 12.0, w2 = 12.0;
-            const double C1 = 1.0 / Mass[0] - A;
-            const double C2 = 1.0 / Mass[1] - A;
-            const double BH_sep2 = (Porg[3] - Porg[0]) * (Porg[3] - Porg[0])
-                                 + (Porg[4] - Porg[1]) * (Porg[4] - Porg[1])
-                                 + (Porg[5] - Porg[2]) * (Porg[5] - Porg[2]);
-            const double inv_BH_sep2 = 1.0 / BH_sep2;
-            for (int k0 = 0; k0 < nz; ++k0) {
-            for (int j0 = 0; j0 < ny; ++j0) {
-            for (int i0 = 0; i0 < nx; ++i0) {
-                const size_t idx = idx_ex(i0, j0, k0, ex);
-                const double xp = X[i0], yp = Y[j0], zp = Z[k0];
-                const double r1 = ((Porg[0]-xp)*(Porg[0]-xp) + (Porg[1]-yp)*(Porg[1]-yp) + (Porg[2]-zp)*(Porg[2]-zp)) * inv_BH_sep2;
-                const double r2 = ((Porg[3]-xp)*(Porg[3]-xp) + (Porg[4]-yp)*(Porg[4]-yp) + (Porg[5]-zp)*(Porg[5]-zp)) * inv_BH_sep2;
-            #if (GAUGE == 6)
-                const double reta_val = A + C1 / (1.0 + w1 * r1) + C2 / (1.0 + w2 * r2);
-            #else
-                const double reta_val = A + C1 * exp(-w1 * r1) + C2 * exp(-w2 * r2);
-            #endif
-                betax_rhs[idx] = FF * dtSfx[idx];
-                betay_rhs[idx] = FF * dtSfy[idx];
-                betaz_rhs[idx] = FF * dtSfz[idx];
-                dtSfx_rhs[idx] = Gamx_rhs[idx] - reta_val * dtSfx[idx];
-                dtSfy_rhs[idx] = Gamy_rhs[idx] - reta_val * dtSfy[idx];
-                dtSfz_rhs[idx] = Gamz_rhs[idx] - reta_val * dtSfz[idx];
-            }}}
-        }
-        else
-        {
-            fprintf(stderr, "z4c_rhs_c: GAUGE %d requires BHN=2, got BHN=%d\n", (int)GAUGE, BHN);
-            return 1;
-        }
-    }
-#else
-#error "z4c_rhs_c.C: unsupported GAUGE value"
-#endif
-
-    lopsided(ex, X, Y, Z, gxx, gxx_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, gxy, gxy_rhs, betax, betay, betaz, Symmetry, AAS);
-    lopsided(ex, X, Y, Z, gxz, gxz_rhs, betax, betay, betaz, Symmetry, ASA);
-    lopsided(ex, X, Y, Z, gyy, gyy_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, gyz, gyz_rhs, betax, betay, betaz, Symmetry, SAA);
-    lopsided(ex, X, Y, Z, gzz, gzz_rhs, betax, betay, betaz, Symmetry, SSS);
-
-    lopsided(ex, X, Y, Z, Axx, Axx_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, Axy, Axy_rhs, betax, betay, betaz, Symmetry, AAS);
-    lopsided(ex, X, Y, Z, Axz, Axz_rhs, betax, betay, betaz, Symmetry, ASA);
-    lopsided(ex, X, Y, Z, Ayy, Ayy_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, Ayz, Ayz_rhs, betax, betay, betaz, Symmetry, SAA);
-    lopsided(ex, X, Y, Z, Azz, Azz_rhs, betax, betay, betaz, Symmetry, SSS);
-
-    lopsided(ex, X, Y, Z, chi_state, chi_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, trK, trK_rhs, betax, betay, betaz, Symmetry, SSS);
-
-    lopsided(ex, X, Y, Z, Gamx, Gamx_rhs, betax, betay, betaz, Symmetry, ASS);
-    lopsided(ex, X, Y, Z, Gamy, Gamy_rhs, betax, betay, betaz, Symmetry, SAS);
-    lopsided(ex, X, Y, Z, Gamz, Gamz_rhs, betax, betay, betaz, Symmetry, SSA);
-
-    lopsided(ex, X, Y, Z, Lap, Lap_rhs, betax, betay, betaz, Symmetry, SSS);
-    lopsided(ex, X, Y, Z, betax, betax_rhs, betax, betay, betaz, Symmetry, ASS);
-    lopsided(ex, X, Y, Z, betay, betay_rhs, betax, betay, betaz, Symmetry, SAS);
-    lopsided(ex, X, Y, Z, betaz, betaz_rhs, betax, betay, betaz, Symmetry, SSA);
-#if (GAUGE == 0 || GAUGE == 2 || GAUGE == 3 || GAUGE == 6 || GAUGE == 7)
-    lopsided(ex, X, Y, Z, dtSfx, dtSfx_rhs, betax, betay, betaz, Symmetry, ASS);
-    lopsided(ex, X, Y, Z, dtSfy, dtSfy_rhs, betax, betay, betaz, Symmetry, SAS);
-    lopsided(ex, X, Y, Z, dtSfz, dtSfz_rhs, betax, betay, betaz, Symmetry, SSA);
-#endif
-    lopsided(ex, X, Y, Z, TZ, TZ_rhs, betax, betay, betaz, Symmetry, SSS);
-
-    for (int idx = 0; idx < all; ++idx)
-    {
-        double Gamxa = 0.0, Gamya = 0.0, Gamza = 0.0;
-        z4c_contract_gamma(
-            gxx[idx], gxy[idx], gxz[idx], gyy[idx], gyz[idx], gzz[idx],
-            gxxx[idx], gxyx[idx], gxzx[idx], gyyx[idx], gyzx[idx], gzzx[idx],
-            gxxy[idx], gxyy[idx], gxzy[idx], gyyy[idx], gyzy[idx], gzzy[idx],
-            gxxz[idx], gxyz[idx], gxzz[idx], gyyz[idx], gyzz[idx], gzzz[idx],
-            Gamxa, Gamya, Gamza);
-
-        TZ_rhs[idx] -= alpn1[idx] * (TWO + kappa2) * kappa1 * TZ[idx];
-        trK_rhs[idx] += alpn1[idx] * kappa1 * (ONE - kappa2) * TZ[idx];
-        Gamx_rhs[idx] -= TWO * alpn1[idx] * kappa1 * (Gamx[idx] - Gamxa);
-        Gamy_rhs[idx] -= TWO * alpn1[idx] * kappa1 * (Gamy[idx] - Gamya);
-        Gamz_rhs[idx] -= TWO * alpn1[idx] * kappa1 * (Gamz[idx] - Gamza);
-    }
-
-    if (eps > 0.0)
-    {
-        kodis(ex, X, Y, Z, chi_state, chi_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, trK, trK_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, gxx, gxx_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, gxy, gxy_rhs, AAS, Symmetry, eps);
-        kodis(ex, X, Y, Z, gxz, gxz_rhs, ASA, Symmetry, eps);
-        kodis(ex, X, Y, Z, gyy, gyy_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, gyz, gyz_rhs, SAA, Symmetry, eps);
-        kodis(ex, X, Y, Z, gzz, gzz_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Axx, Axx_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Axy, Axy_rhs, AAS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Axz, Axz_rhs, ASA, Symmetry, eps);
-        kodis(ex, X, Y, Z, Ayy, Ayy_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Ayz, Ayz_rhs, SAA, Symmetry, eps);
-        kodis(ex, X, Y, Z, Azz, Azz_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Gamx, Gamx_rhs, ASS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Gamy, Gamy_rhs, SAS, Symmetry, eps);
-        kodis(ex, X, Y, Z, Gamz, Gamz_rhs, SSA, Symmetry, eps);
-        kodis(ex, X, Y, Z, Lap, Lap_rhs, SSS, Symmetry, eps);
-        kodis(ex, X, Y, Z, betax, betax_rhs, ASS, Symmetry, eps);
-        kodis(ex, X, Y, Z, betay, betay_rhs, SAS, Symmetry, eps);
-        kodis(ex, X, Y, Z, betaz, betaz_rhs, SSA, Symmetry, eps);
-#if (GAUGE == 0 || GAUGE == 2 || GAUGE == 3 || GAUGE == 6 || GAUGE == 7)
-        kodis(ex, X, Y, Z, dtSfx, dtSfx_rhs, ASS, Symmetry, eps);
-        kodis(ex, X, Y, Z, dtSfy, dtSfy_rhs, SAS, Symmetry, eps);
-        kodis(ex, X, Y, Z, dtSfz, dtSfz_rhs, SSA, Symmetry, eps);
-#endif
-        kodis(ex, X, Y, Z, TZ, TZ_rhs, SSS, Symmetry, eps);
-    }
-
-    if (co == 0)
-    {
-#if (ABV == 0)
-        f_ricci_gamma(ex, X, Y, Z,
-                      chi_constraints,
-                      dxx, gxy, gxz, dyy, gyz, dzz,
-                      Gamx, Gamy, Gamz,
-                      Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-                      Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-                      Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-                      Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-                      Symmetry);
-#endif
-        f_constraint_bssn(ex, X, Y, Z,
-                          chi_constraints, trK,
-                          dxx, gxy, gxz, dyy, gyz, dzz,
-                          Axx, Axy, Axz, Ayy, Ayz, Azz,
-                          Gamx, Gamy, Gamz,
-                          Lap, betax, betay, betaz, rho, Sx, Sy, Sz,
-                          Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-                          Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-                          Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-                          Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-                          Hcon, Mxcon, Mycon, Mzcon, Gmxcon, Gmycon, Gmzcon,
-                          Symmetry);
-    }
-
-    return 0;
-}
-
-extern "C" int f_compute_rhs_Z4c(int *ex, double &T,
-                                 double *X, double *Y, double *Z,
-                                 double *chi, double *trK,
-                                 double *dxx, double *gxy, double *gxz, double *dyy, double *gyz, double *dzz,
-                                 double *Axx, double *Axy, double *Axz, double *Ayy, double *Ayz, double *Azz,
-                                 double *Gamx, double *Gamy, double *Gamz,
-                                 double *Lap, double *betax, double *betay, double *betaz,
-                                 double *dtSfx, double *dtSfy, double *dtSfz,
-                                 double *TZ,
-                                 double *chi_rhs, double *trK_rhs,
-                                 double *gxx_rhs, double *gxy_rhs, double *gxz_rhs, double *gyy_rhs, double *gyz_rhs, double *gzz_rhs,
-                                 double *Axx_rhs, double *Axy_rhs, double *Axz_rhs, double *Ayy_rhs, double *Ayz_rhs, double *Azz_rhs,
-                                 double *Gamx_rhs, double *Gamy_rhs, double *Gamz_rhs,
-                                 double *Lap_rhs, double *betax_rhs, double *betay_rhs, double *betaz_rhs,
-                                 double *dtSfx_rhs, double *dtSfy_rhs, double *dtSfz_rhs,
-                                 double *TZ_rhs,
-                                 double *rho, double *Sx, double *Sy, double *Sz,
-                                 double *Sxx, double *Sxy, double *Sxz, double *Syy, double *Syz, double *Szz,
-                                 double *Gamxxx, double *Gamxxy, double *Gamxxz, double *Gamxyy, double *Gamxyz, double *Gamxzz,
-                                 double *Gamyxx, double *Gamyxy, double *Gamyxz, double *Gamyyy, double *Gamyyz, double *Gamyzz,
-                                 double *Gamzxx, double *Gamzxy, double *Gamzxz, double *Gamzyy, double *Gamzyz, double *Gamzzz,
-                                 double *Rxx, double *Rxy, double *Rxz, double *Ryy, double *Ryz, double *Rzz,
-                                 double *Hcon, double *Mxcon, double *Mycon, double *Mzcon, double *Gmxcon, double *Gmycon, double *Gmzcon,
-                                 int &Symmetry, int &Lev, double &eps, int &co)
-{
-    return compute_rhs_z4c_cartesian(
-        ex, T, X, Y, Z,
-        chi, chi, trK,
-        dxx, gxy, gxz, dyy, gyz, dzz,
-        Axx, Axy, Axz, Ayy, Ayz, Azz,
-        Gamx, Gamy, Gamz,
-        Lap, betax, betay, betaz,
-        dtSfx, dtSfy, dtSfz,
-        TZ,
-        chi_rhs, trK_rhs,
-        gxx_rhs, gxy_rhs, gxz_rhs, gyy_rhs, gyz_rhs, gzz_rhs,
-        Axx_rhs, Axy_rhs, Axz_rhs, Ayy_rhs, Ayz_rhs, Azz_rhs,
-        Gamx_rhs, Gamy_rhs, Gamz_rhs,
-        Lap_rhs, betax_rhs, betay_rhs, betaz_rhs,
-        dtSfx_rhs, dtSfy_rhs, dtSfz_rhs,
-        TZ_rhs,
-        rho, Sx, Sy, Sz,
-        Sxx, Sxy, Sxz, Syy, Syz, Szz,
-        Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-        Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-        Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-        Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-        Hcon, Mxcon, Mycon, Mzcon, Gmxcon, Gmycon, Gmzcon,
-        Symmetry, Lev, eps, co);
-}
-
-extern "C" int f_compute_rhs_Z4cnot(int *ex, double &T,
-                                    double *X, double *Y, double *Z,
-                                    double *chi, double *trK,
-                                    double *dxx, double *gxy, double *gxz, double *dyy, double *gyz, double *dzz,
-                                    double *Axx, double *Axy, double *Axz, double *Ayy, double *Ayz, double *Azz,
-                                    double *Gamx, double *Gamy, double *Gamz,
-                                    double *Lap, double *betax, double *betay, double *betaz,
-                                    double *dtSfx, double *dtSfy, double *dtSfz,
-                                    double *TZ,
-                                    double *chi_rhs, double *trK_rhs,
-                                    double *gxx_rhs, double *gxy_rhs, double *gxz_rhs, double *gyy_rhs, double *gyz_rhs, double *gzz_rhs,
-                                    double *Axx_rhs, double *Axy_rhs, double *Axz_rhs, double *Ayy_rhs, double *Ayz_rhs, double *Azz_rhs,
-                                    double *Gamx_rhs, double *Gamy_rhs, double *Gamz_rhs,
-                                    double *Lap_rhs, double *betax_rhs, double *betay_rhs, double *betaz_rhs,
-                                    double *dtSfx_rhs, double *dtSfy_rhs, double *dtSfz_rhs,
-                                    double *TZ_rhs,
-                                    double *rho, double *Sx, double *Sy, double *Sz,
-                                    double *Sxx, double *Sxy, double *Sxz, double *Syy, double *Syz, double *Szz,
-                                    double *Gamxxx, double *Gamxxy, double *Gamxxz, double *Gamxyy, double *Gamxyz, double *Gamxzz,
-                                    double *Gamyxx, double *Gamyxy, double *Gamyxz, double *Gamyyy, double *Gamyyz, double *Gamyzz,
-                                    double *Gamzxx, double *Gamzxy, double *Gamzxz, double *Gamzyy, double *Gamzyz, double *Gamzzz,
-                                    double *Rxx, double *Rxy, double *Rxz, double *Ryy, double *Ryz, double *Rzz,
-                                    double *Hcon, double *Mxcon, double *Mycon, double *Mzcon, double *Gmxcon, double *Gmycon, double *Gmzcon,
-                                    int &Symmetry, int &Lev, double &eps, int &co, double &chitiny)
-{
-    const int all = ex[0] * ex[1] * ex[2];
-    std::vector<double> chi_clamped(chi, chi + all);
-    f_lowerboundset(ex, chi_clamped.data(), chitiny);
-
-    const int ret = compute_rhs_z4c_cartesian(
-        ex, T, X, Y, Z,
-        chi_clamped.data(), chi, trK,
-        dxx, gxy, gxz, dyy, gyz, dzz,
-        Axx, Axy, Axz, Ayy, Ayz, Azz,
-        Gamx, Gamy, Gamz,
-        Lap, betax, betay, betaz,
-        dtSfx, dtSfy, dtSfz,
-        TZ,
-        chi_rhs, trK_rhs,
-        gxx_rhs, gxy_rhs, gxz_rhs, gyy_rhs, gyz_rhs, gzz_rhs,
-        Axx_rhs, Axy_rhs, Axz_rhs, Ayy_rhs, Ayz_rhs, Azz_rhs,
-        Gamx_rhs, Gamy_rhs, Gamz_rhs,
-        Lap_rhs, betax_rhs, betay_rhs, betaz_rhs,
-        dtSfx_rhs, dtSfy_rhs, dtSfz_rhs,
-        TZ_rhs,
-        rho, Sx, Sy, Sz,
-        Sxx, Sxy, Sxz, Syy, Syz, Szz,
-        Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-        Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-        Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-        Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-        Hcon, Mxcon, Mycon, Mzcon, Gmxcon, Gmycon, Gmzcon,
-        Symmetry, Lev, eps, co);
-
-    if (ret != 0 || co != 0)
-        return ret;
-
-#if (ABV == 0)
-    f_ricci_gamma(ex, X, Y, Z,
-                  chi,
-                  dxx, gxy, gxz, dyy, gyz, dzz,
-                  Gamx, Gamy, Gamz,
-                  Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-                  Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-                  Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-                  Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-                  Symmetry);
-#endif
-    f_constraint_bssn(ex, X, Y, Z,
-                      chi, trK,
-                      dxx, gxy, gxz, dyy, gyz, dzz,
-                      Axx, Axy, Axz, Ayy, Ayz, Azz,
-                      Gamx, Gamy, Gamz,
-                      Lap, betax, betay, betaz, rho, Sx, Sy, Sz,
-                      Gamxxx, Gamxxy, Gamxxz, Gamxyy, Gamxyz, Gamxzz,
-                      Gamyxx, Gamyxy, Gamyxz, Gamyyy, Gamyyz, Gamyzz,
-                      Gamzxx, Gamzxy, Gamzxz, Gamzyy, Gamzyz, Gamzzz,
-                      Rxx, Rxy, Rxz, Ryy, Ryz, Rzz,
-                      Hcon, Mxcon, Mycon, Mzcon, Gmxcon, Gmycon, Gmzcon,
-                      Symmetry);
-    return ret;
-}

BSSN_BUILD_CONFIG_MIGRATION.md

@@ -1,211 +0,0 @@
-# BSSN Build Config Migration
-
-This note records the build-configuration fix needed when replacing
-`AMSS_NCKU_Input.py` or `generate_macrodef.py` with a newer upstream version.
-
-## Problem
-
-`AMSS_NCKU_source/macrodef.h` is not the authoritative file used by normal
-runs. `AMSS_NCKU_Program.py` first generates macro files under
-`input_data.File_directory`, copies `AMSS_NCKU_source` to
-`<File_directory>/AMSS_NCKU_source_copy`, then copies the generated macro files
-into that copied source tree and compiles there.
-
-Therefore, makefile logic must not depend only on the stale
-`AMSS_NCKU_source/macrodef.h`. The actual equation path must be passed to the
-copied build tree from the same generation step that creates `macrodef.h`.
-
-The performance regression was caused by compiling/linking the
-`BSSN-EScalar` C wrapper into BSSN vacuum builds. For BSSN vacuum (`ABEtype=0`),
-the build must use:
-
-```make
-BSSN_USE_TRANSFER_CACHE=1
-BSSN_USE_ESCALAR_C_KERNEL=0
-```
-
-and must not link `bssn_escalar_rhs_c.o`.
-
-## Required Migration Steps
-
-### 1. Add an ABE type helper in `generate_macrodef.py`
-
-Add a helper that maps `input_data.Equation_Class` to the numeric `ABEtype`.
-Use the same mapping as `macrodef.h`:
-
-```python
-def get_abe_type():
-    if ( input_data.Equation_Class == "BSSN" ):
-        return 0
-    elif ( input_data.Equation_Class == "BSSN-EScalar" ):
-        return 1
-    elif ( input_data.Equation_Class == "BSSN-EM" ):
-        return 3
-    elif ( input_data.Equation_Class == "Z4C" ):
-        return 2
-    else:
-        raise ValueError("Equation_Class setting error!!!")
-```
-
-Update `generate_macrodef_h()` to print `#define ABEtype {get_abe_type()}`
-instead of duplicating the if/elif mapping.
-
-### 2. Generate a makefile fragment
-
-In `generate_macrodef.py`, add:
-
-```python
-def generate_build_config():
-    file1 = open(os.path.join(input_data.File_directory, "AMSS_NCKU_build.mk"), "w")
-    print("# Generated by generate_macrodef.py; do not edit manually.", file=file1)
-    print(f"ABE_TYPE := {get_abe_type()}", file=file1)
-    file1.close()
-```
-
-This file is the build-time authority for the equation path.
-
-### 3. Call and copy the generated build config
-
-In `AMSS_NCKU_Program.py`, after generating `macrodef.h` and `macrodef.fh`, call:
-
-```python
-generate_macrodef.generate_build_config()
-print(" AMSS-NCKU build config AMSS_NCKU_build.mk has been generated. ")
-```
-
-When copying generated files into `AMSS_NCKU_source_copy`, also copy:
-
-```python
-build_config_path = os.path.join(File_directory, "AMSS_NCKU_build.mk")
-shutil.copy2(build_config_path, AMSS_NCKU_source_copy)
-```
-
-### 4. Make the source makefile consume the generated config
-
-At the top of `AMSS_NCKU_source/makefile`, after `include makefile.inc`, add:
-
-```make
--include AMSS_NCKU_build.mk
-
-ABE_TYPE ?= $(shell awk '/^[[:space:]]*\#define[[:space:]]+ABEtype/ {print $$3; exit}' macrodef.h 2>/dev/null)
-```
-
-The generated `AMSS_NCKU_build.mk` is used during normal Python-driven builds.
-The fallback keeps manual source-tree builds usable.
-
-### 5. Gate path-specific build options by `ABE_TYPE`
-
-Use effective build switches:
-
-```make
-ifeq ($(USE_TRANSFER_CACHE),auto)
-ifeq ($(ABE_TYPE),0)
-EFFECTIVE_USE_TRANSFER_CACHE = 1
-else
-EFFECTIVE_USE_TRANSFER_CACHE = 0
-endif
-else
-EFFECTIVE_USE_TRANSFER_CACHE = $(USE_TRANSFER_CACHE)
-endif
-
-ifeq ($(USE_CXX_ESCALAR_KERNEL),1)
-ifeq ($(ABE_TYPE),1)
-EFFECTIVE_USE_CXX_ESCALAR_KERNEL = 1
-else
-EFFECTIVE_USE_CXX_ESCALAR_KERNEL = 0
-endif
-else
-EFFECTIVE_USE_CXX_ESCALAR_KERNEL = 0
-endif
-
-TRANSFER_CACHE_FLAG = -DBSSN_USE_TRANSFER_CACHE=$(EFFECTIVE_USE_TRANSFER_CACHE)
-ESCALAR_KERNEL_FLAG = -DBSSN_USE_ESCALAR_C_KERNEL=$(EFFECTIVE_USE_CXX_ESCALAR_KERNEL)
-```
-
-Only add `bssn_escalar_rhs_c.o` when the effective EScalar C kernel switch is
-enabled:
-
-```make
-ifeq ($(EFFECTIVE_USE_CXX_ESCALAR_KERNEL),1)
-CFILES += bssn_escalar_rhs_c.o
-endif
-```
-
-### 6. Use safe transfer-cache default
-
-In `AMSS_NCKU_source/makefile.inc`, keep:
-
-```make
-USE_TRANSFER_CACHE ?= auto
-```
-
-With the effective switch logic above, this enables cached transfer for BSSN
-vacuum while keeping non-BSSN paths on the uncached path by default.
-
-## Verification Checklist
-
-Run these checks after migrating:
-
-```bash
-python3 -c "import generate_macrodef; generate_macrodef.generate_build_config()"
-cat GW150914/AMSS_NCKU_build.mk
-```
-
-For BSSN, the generated file should contain:
-
-```make
-ABE_TYPE := 0
-```
-
-Dry-run the copied or source makefile:
-
-```bash
-make -n -B INTERP_LB_MODE=off ABE | grep -E 'BSSN_USE_TRANSFER_CACHE|BSSN_USE_ESCALAR_C_KERNEL|bssn_escalar_rhs_c'
-```
-
-Expected BSSN result:
-
-```text
--DBSSN_USE_TRANSFER_CACHE=1 -DBSSN_USE_ESCALAR_C_KERNEL=0
-```
-
-and no `bssn_escalar_rhs_c.o` in the final link command.
-
-Run the full workflow:
-
-```bash
-python3 AMSS_NCKU_Program.py
-```
-
-For the 10-step BSSN test, compare coordinate output:
-
-```bash
-python3 - <<'PY'
-from pathlib import Path
-old = Path('../GW150914-06457/AMSS_NCKU_output/bssn_BH.dat')
-new = Path('GW150914/AMSS_NCKU_output/bssn_BH.dat')
-
-def rows(path):
-    out = []
-    for line in path.read_text().splitlines():
-        if not line.strip() or line.lstrip().startswith('#'):
-            continue
-        out.append([float(x) for x in line.split()])
-    return out
-
-ro, rn = rows(old), rows(new)
-n = min(len(ro), len(rn))
-max_abs = 0.0
-for i in range(n):
-    for a, b in zip(ro[i], rn[i]):
-        max_abs = max(max_abs, abs(a - b))
-print(f"old_rows={len(ro)} new_rows={len(rn)} compared_rows={n}")
-print(f"max_abs_diff={max_abs:.17g}")
-PY
-```
-
-For the validated migration, the first 10 rows matched exactly:
-
-```text
-max_abs_diff=0
-```

generate_macrodef.py

@@ -12,37 +12,6 @@ import os
 import AMSS_NCKU_Input as input_data          ## import program input file
 
 
-##################################################################
-
-def get_abe_type():
-    if ( input_data.Equation_Class == "BSSN" ):
-        return 0
-    elif ( input_data.Equation_Class == "BSSN-EScalar" ):
-        return 1
-    elif ( input_data.Equation_Class == "BSSN-EM" ):
-        return 3
-    elif ( input_data.Equation_Class == "Z4C" ):
-        return 2
-    else:
-        raise ValueError("Equation_Class setting error!!!")
-
-
-##################################################################
-
-## Generate the makefile fragment used by the copied source tree.
-## The source-tree macrodef.h is not authoritative because macro files
-## are regenerated under File_directory for each run.
-
-def generate_build_config():
-
-    file1 = open( os.path.join(input_data.File_directory, "AMSS_NCKU_build.mk"), "w")
-
-    print( "# Generated by generate_macrodef.py; do not edit manually.", file=file1 )
-    print( f"ABE_TYPE := {get_abe_type()}",                             file=file1 )
-
-    file1.close()
-
-
 ##################################################################
 
 ## Generate the macro file macrodef.h according to user settings
@@ -89,10 +58,19 @@ def generate_macrodef_h():
     # 2: Z4c vacuum
     # 3: coupled to Maxwell field
 
-    try:
-        print( f"#define ABEtype {get_abe_type()}", file=file1 )
-        print(                                      file=file1 )
-    except ValueError:
+    if ( input_data.Equation_Class == "BSSN" ):
+        print( "#define ABEtype 0", file=file1 )
+        print(                      file=file1 )
+    elif ( input_data.Equation_Class == "BSSN-EScalar" ):
+        print( "#define ABEtype 1", file=file1 )
+        print(                      file=file1 )
+    elif ( input_data.Equation_Class == "BSSN-EM" ):
+        print( "#define ABEtype 3", file=file1 )
+        print(                      file=file1 )
+    elif ( input_data.Equation_Class == "Z4C" ):
+        print( "#define ABEtype 2", file=file1 )
+        print(                      file=file1 )
+    else:
         print( "Equation_Class setting error!!!"                )
         print()
         print( "# Equation type #define ABEtype setting error!!!", file=file1 )
@@ -166,62 +144,6 @@ def generate_macrodef_h():
     print( "#define REGLEV 0",      file=file1 )
     print(                          file=file1 )
 
-    # Define fine-grained timing/debug macros.
-    # All of them default to OFF so production builds do not pay profiling overhead.
-
-    fine_timing = getattr(input_data, "Fine_Timing",
-                  getattr(input_data, "Finegrained_Timing", "no"))
-    kernel_fine_timing = getattr(input_data, "Kernel_Fine_Timing",
-                          getattr(input_data, "BSSN_Kernel_Fine_Timing", "no"))
-    stdin_abort_poll = getattr(input_data, "Enable_Stdin_Abort_Poll",
-                       getattr(input_data, "Stdin_Abort_Poll", "no"))
-    timing_report_every = max(1, int(getattr(
-        input_data, "Timing_Every_Steps",
-        getattr(input_data, "Timing_Report_Every", 1))))
-    timing_top_hotspots = max(1, int(getattr(
-        input_data, "Timing_Top_Hotspots", 8)))
-
-    if ( fine_timing == "yes" ):
-        print( "#define BSSN_FINE_TIMING 1", file=file1 )
-        print(                               file=file1 )
-    elif ( fine_timing == "no" ):
-        print( "#define BSSN_FINE_TIMING 0", file=file1 )
-        print(                               file=file1 )
-    else:
-        print( "Fine_Timing setting error!!!" )
-        print()
-        print( "# Fine_Timing setting error!!!", file=file1 )
-        print(                                   file=file1 )
-
-    print( f"#define BSSN_FINE_TIMING_EVERY {timing_report_every}", file=file1 )
-    print(                                                          file=file1 )
-    print( f"#define BSSN_FINE_TIMING_TOPN {timing_top_hotspots}",  file=file1 )
-    print(                                                          file=file1 )
-
-    if ( kernel_fine_timing == "yes" ):
-        print( "#define BSSN_KERNEL_FINE_TIMING 1", file=file1 )
-        print(                                      file=file1 )
-    elif ( kernel_fine_timing == "no" ):
-        print( "#define BSSN_KERNEL_FINE_TIMING 0", file=file1 )
-        print(                                      file=file1 )
-    else:
-        print( "Kernel_Fine_Timing setting error!!!" )
-        print()
-        print( "# Kernel_Fine_Timing setting error!!!", file=file1 )
-        print(                                          file=file1 )
-
-    if ( stdin_abort_poll == "yes" ):
-        print( "#define BSSN_ENABLE_STDIN_ABORT_POLL 1", file=file1 )
-        print(                                           file=file1 )
-    elif ( stdin_abort_poll == "no" ):
-        print( "#define BSSN_ENABLE_STDIN_ABORT_POLL 0", file=file1 )
-        print(                                           file=file1 )
-    else:
-        print( "Enable_Stdin_Abort_Poll setting error!!!" )
-        print()
-        print( "# Enable_Stdin_Abort_Poll setting error!!!", file=file1 )
-        print(                                               file=file1 )
-
     # Define macro USE_GPU
     # use GPU or not
     
@@ -302,21 +224,6 @@ def generate_macrodef_h():
     print( "//     0: for every level;",                                                 file=file1 ) 
     print( "//     1: for all",                                                          file=file1 )
     print( "//",                                                                         file=file1 )
-    print( "// define BSSN_FINE_TIMING",                                                 file=file1 )
-    print( "//     enable fine-grained per-timestep timing monitor",                     file=file1 )
-    print( "//",                                                                         file=file1 )
-    print( "// define BSSN_FINE_TIMING_EVERY",                                           file=file1 )
-    print( "//     report timing every N coarse timesteps",                              file=file1 )
-    print( "//",                                                                         file=file1 )
-    print( "// define BSSN_FINE_TIMING_TOPN",                                            file=file1 )
-    print( "//     number of hottest timing buckets shown in stdout",                    file=file1 )
-    print( "//",                                                                         file=file1 )
-    print( "// define BSSN_KERNEL_FINE_TIMING",                                          file=file1 )
-    print( "//     enable split timing inside compute_rhs_bssn",                         file=file1 )
-    print( "//",                                                                         file=file1 )
-    print( "// define BSSN_ENABLE_STDIN_ABORT_POLL",                                     file=file1 )
-    print( "//     poll stdin and broadcast abort flag every coarse step",               file=file1 )
-    print( "//",                                                                         file=file1 )
     print( "// define USE_GPU",                                                          file=file1 )
     print( "//     use gpu or not",                                                      file=file1 )
     print( "//",                                                                         file=file1 )

makefile_and_run.py

@@ -9,6 +9,7 @@
 
 
 import AMSS_NCKU_Input as input_data
+import os
 import subprocess
 import time
 
@@ -57,6 +58,48 @@ BUILD_JOBS = 64
 ##################################################################
 
 
+##################################################################
+
+def prepare_gpu_runtime_env():
+    """
+    Create a user-private CUDA MPS environment for GPU runs.
+
+    On shared machines another user's daemon may already occupy the default
+    /tmp/nvidia-mps pipe directory, which makes plain cudaSetDevice/cudaMalloc
+    fail with cudaErrorMpsConnectionFailed.  Binding AMSS-NCKU to a private
+    pipe directory avoids cross-user interference.
+    """
+    env = os.environ.copy()
+
+    pipe_dir = env.get("CUDA_MPS_PIPE_DIRECTORY", f"/tmp/amss-ncku-mps-{os.getuid()}")
+    log_dir  = env.get("CUDA_MPS_LOG_DIRECTORY",  f"/tmp/amss-ncku-mps-log-{os.getuid()}")
+
+    os.makedirs(pipe_dir, exist_ok=True)
+    os.makedirs(log_dir, exist_ok=True)
+
+    env["CUDA_MPS_PIPE_DIRECTORY"] = pipe_dir
+    env["CUDA_MPS_LOG_DIRECTORY"] = log_dir
+
+    control_socket = os.path.join(pipe_dir, "control")
+    if not os.path.exists(control_socket):
+        start = subprocess.run(
+            ["nvidia-cuda-mps-control", "-d"],
+            env=env,
+            stdout=subprocess.DEVNULL,
+            stderr=subprocess.DEVNULL,
+        )
+        if start.returncode != 0:
+            print(f" Warning: failed to start private CUDA MPS daemon in {pipe_dir}")
+        else:
+            print(f" Using private CUDA MPS pipe directory: {pipe_dir}")
+    else:
+        print(f" Using existing private CUDA MPS pipe directory: {pipe_dir}")
+
+    return env
+
+##################################################################
+
+
 
 ##################################################################
 
@@ -146,16 +189,29 @@ def run_ABE():
 
     ## Define the command to run; cast other values to strings as needed
     
+    run_env = None
+
     if (input_data.GPU_Calculation == "no"):
         mpi_command         = NUMACTL_CPU_BIND + " mpirun -np " + str(input_data.MPI_processes) + " ./ABE"
         #mpi_command         = " mpirun -np " + str(input_data.MPI_processes) + " ./ABE"
         mpi_command_outfile = "ABE_out.log"
     elif (input_data.GPU_Calculation == "yes"):
-        mpi_command         = NUMACTL_CPU_BIND + " mpirun -np " + str(input_data.MPI_processes) + " ./ABEGPU"
+        run_env = prepare_gpu_runtime_env()
+        if int(input_data.MPI_processes) == 1:
+            mpi_command = "./ABEGPU"
+        else:
+            mpi_command = NUMACTL_CPU_BIND + " mpirun -np " + str(input_data.MPI_processes) + " ./ABEGPU"
         mpi_command_outfile = "ABEGPU_out.log"
  
     ## Execute the MPI command and stream output
-    mpi_process = subprocess.Popen(mpi_command, shell=True, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, text=True)
+    mpi_process = subprocess.Popen(
+        mpi_command,
+        shell=True,
+        stdout=subprocess.PIPE,
+        stderr=subprocess.STDOUT,
+        text=True,
+        env=run_env,
+    )
 
     ## Write ABE run output to file while printing to stdout
     with open(mpi_command_outfile, 'w') as file0: