Commit 875844ff authored by Elwasif, Wael's avatar Elwasif, Wael
Browse files

Resolved merge conflict by incorporating both suggestions.

parents 53611d2f eca5ca7c
Loading
Loading
Loading
Loading

src/Abatch.00000.m

0 → 100644
+19115 −0

File added.

Preview size limit exceeded, changes collapsed.

src/BatchedGemm.o

0 → 100644
+3.45 KiB

File added.

No diff preview for this file type.

src/Bbatch.00000.m

0 → 100644
+19115 −0

File added.

Preview size limit exceeded, changes collapsed.

src/all_analysis.csv

0 → 100644
+8 −0
Original line number Diff line number Diff line
Instance, 
0,   
1,   
2,   
3,   
4,   
5,   
6,   
+9 −4
Original line number Diff line number Diff line
@@ -26,12 +26,15 @@ void apply_Htarget_pvbatch(
#define Y(i) Y_[(i)-1]
{
 const int ialign = 32;
 const double giga = 1000.0 * 1000.0 * 1000.0;

 double gflops1 = (double) 0.0;
 double gflops2 = (double) 0.0;
 double time_1st_vbatch = (double) 0.0;
 double time_2nd_vbatch = (double) 0.0;

 size_t nbytes_BX = 0;

/*
 ------------------
 compute  Y = H * X
@@ -141,7 +144,8 @@ void apply_Htarget_pvbatch(
 int ld_Abatch = descAbatch_[LLD_];
 int ld_Bbatch = descBbatch_[LLD_];

 double *BX_ = (double *) dmrg_malloc( (sizeof(double) * ld_BX) * Locq_BX );
 nbytes_BX = ( (sizeof(double) * ld_BX) * Locq_BX );
 double *BX_ = (double *) dmrg_malloc( nbytes_BX );
 assert( BX_ != NULL );

#define BX(i,j) BX_[ indx2f(i,j,ld_BX) ]
@@ -288,7 +292,7 @@ void apply_Htarget_pvbatch(
#ifdef _OPENMP
    time_1st_vbatch += omp_get_wtime();
#endif
    gflops1 = gflops1/(1000.0*1000.0*1000.0);
    gflops1 = gflops1/(giga);
   


@@ -363,16 +367,17 @@ void apply_Htarget_pvbatch(
                      ngroups, group_size_ );
#ifdef _OPENMP
   time_2nd_vbatch += omp_get_wtime();
   gflops2 = gflops2/(1000.0*1000.0*1000.0);
   gflops2 = gflops2/(giga);
   time_t sec;
   long  ms;
   get_current_time_with_ms (&sec , &ms);

   printf("1st vbatch %f gflops (gflops1=%lf,time=%lf)\n", 
          gflops1/time_1st_vbatch,  gflops1, time_1st_vbatch );
   printf("2nd vbatch %f gflops (gflops2=%lf,time=%lf)\n", 
          gflops2/time_2nd_vbatch, gflops2, time_2nd_vbatch );

   printf(" %"PRIdMAX".%03ld : overall %f gflops\n", (intmax_t)s, ms, (gflops1+gflops2)/(time_1st_vbatch + time_2nd_vbatch) );
   printf(" %"PRIdMAX".%03ld : overall %f gflops memory BX (%f GBytes)\n", (intmax_t)s, ms, (gflops1+gflops2)/(time_1st_vbatch + time_2nd_vbatch), (double) nbytes_BX/(giga) );
#endif
     

Loading