Optimize GPU RK4 stage sync path

AMSS_NCKU_source/ABE.C

@@ -23,7 +23,10 @@ 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"
@@ -469,10 +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;