portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-04-10 06:00:45 +01:00
Code Releases Activity

work distribution

Browse Source
This commit is contained in:
Peter Boyle 2018-04-12 11:41:41 -04:00
parent 7533035a99
commit 6411caad67
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
lib/threads/Pragmas.h
View File

@ -77,7 +77,7 @@ Author: paboyle <paboyle@ph.ed.ac.uk>
////////////////////////////////////////////////////////////////////////////////// //////////////////////////////////////////////////////////////////////////////////
#ifdef GRID_NVCC #ifdef GRID_NVCC
constexpr uint32_t gpu_threads = 8; constexpr uint32_t gpu_threads = 32;
template<typename lambda> __global__ template<typename lambda> __global__
void LambdaApply(uint64_t base, uint64_t Num, lambda Lambda) void LambdaApply(uint64_t base, uint64_t Num, lambda Lambda)
@ -98,7 +98,7 @@ void LambdaApply(uint64_t base, uint64_t Num, lambda Lambda)
Iterator num = range.end(); \ Iterator num = range.end(); \
Iterator base = range.begin(); \ Iterator base = range.begin(); \
Iterator num_block = (num + gpu_threads - 1)/gpu_threads; \ Iterator num_block = (num + gpu_threads - 1)/gpu_threads; \
LambdaApply<<<num_block,gpu_threads>>>(base,num_,lambda); \ LambdaApply<<<num_block,gpu_threads>>>(base,num,lambda); \
cudaDeviceSynchronize(); \ cudaDeviceSynchronize(); \
cudaError err = cudaGetLastError(); \ cudaError err = cudaGetLastError(); \
if ( cudaSuccess != err ) { \ if ( cudaSuccess != err ) { \