portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-12-23 14:14:30 +00:00
Code Releases Activity

Gives 200GF/s on SyCL/DG1 8^4, doesn't uglify develop for other platforms too badly.

Browse Source 
Easy to revert to clean more C++ stylistic code. Theres a SYCL_HACK macro I will clean up later once dpcpp
evolves a central nervous systems.
This commit is contained in:
u61464
2021-03-10 05:40:51 -08:00
parent 2146eebb65
commit 0e21adb3f6
3 changed files with 115 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
Grid/qcd/action/fermion/implementation/WilsonKernelsImplementation.h
View File

@@ -416,7 +416,21 @@ void WilsonKernels<Impl>::DhopDirKernel( StencilImpl &st, DoubledGaugeField &U,S
#undef LoopBody
}
#define KERNEL_CALLNB(A) \
#define KERNEL_CALL_TMP(A) \
const uint64_t NN = Nsite*Ls; \
auto U_p = & U_v[0]; \
auto in_p = & in_v[0]; \
auto out_p = & out_v[0]; \
auto st_p = st_v._entries_p; \
auto st_perm = st_v._permute_type; \
accelerator_forNB( ss, NN, Simd::Nsimd(), { \
int sF = ss; \
int sU = ss/Ls; \
WilsonKernels<Impl>::A(st_perm,st_p,U_p,buf,sF,sU,in_p,out_p); \
}); \
accelerator_barrier();
#define KERNEL_CALLNB(A) \
const uint64_t NN = Nsite*Ls; \
accelerator_forNB( ss, NN, Simd::Nsimd(), { \
int sF = ss; \
@@ -445,7 +459,11 @@ void WilsonKernels<Impl>::DhopKernel(int Opt,StencilImpl &st, DoubledGaugeField
if( interior && exterior ) {
if (Opt == WilsonKernelsStatic::OptGeneric ) { KERNEL_CALL(GenericDhopSite); return;}
#ifdef SYCL_HACK
if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL_TMP(HandDhopSiteSycl); return; }
#else
if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL(HandDhopSite); return;}
#endif
#ifndef GRID_CUDA
if (Opt == WilsonKernelsStatic::OptInlineAsm ) { ASM_CALL(AsmDhopSite); return;}
#endif