portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-04-04 19:25:56 +01:00
Code Releases Activity

Updating

Browse Source
This commit is contained in:
Peter Boyle 2016-12-01 12:42:53 +00:00
parent f7a6b8e5ed
commit e27c6b217c
3 changed files with 10 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
benchmarks/Benchmark_mooee.cc
View File

@ -101,6 +101,7 @@ int main (int argc, char ** argv)
#define BENCH_DW(A,in,out) \ #define BENCH_DW(A,in,out) \
Dw.CayleyZeroCounters(); \ Dw.CayleyZeroCounters(); \
Dw. A (in,out); \
FGrid->Barrier(); \ FGrid->Barrier(); \
t0=usecond(); \ t0=usecond(); \
for(int i=0;i<ncall;i++){ \ for(int i=0;i<ncall;i++){ \
@ -114,6 +115,7 @@ int main (int argc, char ** argv)
#define BENCH_DW_MEO(A,in,out) \ #define BENCH_DW_MEO(A,in,out) \
Dw.CayleyZeroCounters(); \ Dw.CayleyZeroCounters(); \
Dw. A (in,out,0); \
FGrid->Barrier(); \ FGrid->Barrier(); \
t0=usecond(); \ t0=usecond(); \
for(int i=0;i<ncall;i++){ \ for(int i=0;i<ncall;i++){ \

5
lib/qcd/action/fermion/CayleyFermion5Dvec.cc
View File

@ -124,7 +124,10 @@ PARALLEL_FOR_LOOP
} }
#else #else
for(int v=0;v<LLs;v++){ for(int v=0;v<LLs;v++){
vprefetch(psi[ss+v+LLs]);
// vprefetch(phi[ss+v+LLs]);
int vp= (v==LLs-1) ? 0 : v+1; int vp= (v==LLs-1) ? 0 : v+1;
int vm= (v==0 ) ? LLs-1 : v-1; int vm= (v==0 ) ? LLs-1 : v-1;

8
lib/simd/Grid_avx512.h
View File

@ -86,13 +86,13 @@ namespace Optimization {
struct Vstream{ struct Vstream{
//Float //Float
inline void operator()(float * a, __m512 b){ inline void operator()(float * a, __m512 b){
//_mm512_stream_ps(a,b); _mm512_stream_ps(a,b);
_mm512_store_ps(a,b); // _mm512_store_ps(a,b);
} }
//Double //Double
inline void operator()(double * a, __m512d b){ inline void operator()(double * a, __m512d b){
//_mm512_stream_pd(a,b); _mm512_stream_pd(a,b);
_mm512_store_pd(a,b); // _mm512_store_pd(a,b);
} }
}; };