portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-10 07:55:35 +00:00
Code Releases Activity

Dont stream as poor perf in some cases

Browse Source
This commit is contained in:
Peter Boyle 2019-07-01 07:30:25 +01:00
parent 7b7c470917
commit b8f7bfbb26
1 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
Grid/tensors/Tensor_SIMT.h
View File

@ -35,7 +35,8 @@ NAMESPACE_BEGIN(Grid);
accelerator_inline void synchronise(void) accelerator_inline void synchronise(void)
{ {
#ifdef __CUDA_ARCH__ #ifdef __CUDA_ARCH__
__syncthreads(); // __syncthreads();
__syncwarp();
#endif #endif
return; return;
} }
@ -65,7 +66,8 @@ vobj coalescedReadPermute(const vobj & __restrict__ vec,int ptype,int doperm,int
template<class vobj> accelerator_inline template<class vobj> accelerator_inline
void coalescedWrite(vobj & __restrict__ vec,const vobj & __restrict__ extracted,int lane=0) void coalescedWrite(vobj & __restrict__ vec,const vobj & __restrict__ extracted,int lane=0)
{ {
vstream(vec, extracted); // vstream(vec, extracted);
vec = extracted;
} }
#else #else
accelerator_inline int SIMTlane(int Nsimd) { return threadIdx.y; } // CUDA specific accelerator_inline int SIMTlane(int Nsimd) { return threadIdx.y; } // CUDA specific