moebiusband/MD-Bench
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Use aligned loads when gathering j atoms

Browse Source 
Signed-off-by: Rafael Ravedutti <rafaelravedutti@gmail.com>
This commit is contained in:
Rafael Ravedutti 2022-02-04 14:29:32 +01:00
parent 7b90800a2b
commit 6e6a3f6502
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
gromacs/includes/simd.h
View File

@ -51,8 +51,8 @@ static MD_SIMD_FLOAT simd_gather2(MD_FLOAT *c0, MD_FLOAT *c1, int d) {
x = _mm512_mask_i32gather_pd(simd_zero(), simd_mask_from_u32(0x0f), vindex, c0, sizeof(double)); x = _mm512_mask_i32gather_pd(simd_zero(), simd_mask_from_u32(0x0f), vindex, c0, sizeof(double));
x = _mm512_mask_i32gather_pd(x, simd_mask_from_u32(0xf0), vindex, c1, sizeof(double)); x = _mm512_mask_i32gather_pd(x, simd_mask_from_u32(0xf0), vindex, c1, sizeof(double));
#else #else
x = _mm512_loadu_pd(&c0[d * CLUSTER_DIM_M]); x = _mm512_load_pd(&c0[d * CLUSTER_DIM_M]);
x = _mm512_insertf64x4(x, _mm256_loadu_pd(&c1[d * CLUSTER_DIM_M]), 1); x = _mm512_insertf64x4(x, _mm256_load_pd(&c1[d * CLUSTER_DIM_M]), 1);
#endif #endif
return x; return x;
} }