Add thread-safe ShellPatch::setupintintstuff with OpenMP

Split prolongpointstru into search-only (prolongpointstru_search) and
append-only (prolongpointstru_append) functions. The search is read-only
and thread-safe; each thread builds private linked lists via
prolongpointstru_append, merged after the parallel loop.

This eliminates critical-section contention and delivers ~2.2x speedup:
setupintintstuff: 511s -> 252s, total init: 592s -> 267s.

Also add -qopenmp to ShellPatch.o compilation via makefile override rule
and <omp.h> include with _OPENMP guards + fallback stubs.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

AMSS_NCKU_source/ShellPatch.h

@@ -102,6 +102,16 @@ 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
@@ -175,6 +185,12 @@ 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 */,
@@ -195,11 +211,11 @@ public:
    bool Interp_One_Point(MyList<var> *VarList,
                          double *XX, /*input global Cartesian coordinate*/
                          double *Shellf, int Symmetry);
-   void write_Pablo_file_ss(int *ext, double xmin, double xmax, double ymin, double ymax, double zmin, double zmax,
+   void write_Pablo_file_ss(int *ext, double xmin, double xmax, double ymin, double ymax, double zmin, double zmax,
-                            char *filename, int sst);
+                            char *filename, int sst);
-   double L2Norm(var *vf);
+   double L2Norm(var *vf);
-   void L2Norm7(var **vf, double *norms);
+   void L2Norm7(var **vf, double *norms);
-   void Find_Maximum(MyList<var> *VarList, double *XX, double *Shellf);
+   void Find_Maximum(MyList<var> *VarList, double *XX, double *Shellf);
-};
+};
 
 #endif /* SHELLPATCH_H */

AMSS_NCKU_source/makefile

@@ -56,6 +56,10 @@ endif
 .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 $@