portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-11-04 05:54:32 +00:00
Code Releases Activity

Threading support rework.

Browse Source 
Placed parallel pragmas as macros; implemented deterministic thread reduction in style of
BFM.
This commit is contained in:
Peter Boyle
2015-05-12 07:51:41 +01:00
parent 6e6843ac69
commit c6baa3e657
26 changed files with 276 additions and 184 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
lib/cshift/Grid_cshift_common.h
View File

@@ -28,7 +28,7 @@ Gather_plane_simple (const Lattice<vobj> &rhs,std::vector<cobj,alignedAllocator<
int so = plane*rhs._grid->_ostride[dimension]; // base offset for start of plane
int bo = 0; // offset in buffer
#pragma omp parallel for collapse(2)
PARALLEL_NESTED_LOOP(2)
for(int n=0;n<rhs._grid->_slice_nblock[dimension];n++){
for(int b=0;b<rhs._grid->_slice_block[dimension];b++){
int o = n*rhs._grid->_slice_stride[dimension];
@@ -57,7 +57,7 @@ Gather_plane_extract(const Lattice<vobj> &rhs,std::vector<typename cobj::scalar_
int so = plane*rhs._grid->_ostride[dimension]; // base offset for start of plane
int bo = 0; // offset in buffer
#pragma omp parallel for collapse(2)
PARALLEL_NESTED_LOOP(2)
for(int n=0;n<rhs._grid->_slice_nblock[dimension];n++){
for(int b=0;b<rhs._grid->_slice_block[dimension];b++){
@@ -106,7 +106,7 @@ template<class vobj> void Scatter_plane_simple (Lattice<vobj> &rhs,std::vector<v
int so = plane*rhs._grid->_ostride[dimension]; // base offset for start of plane
int bo = 0; // offset in buffer
#pragma omp parallel for collapse(2)
PARALLEL_NESTED_LOOP(2)
for(int n=0;n<rhs._grid->_slice_nblock[dimension];n++){
for(int b=0;b<rhs._grid->_slice_block[dimension];b++){
int o=n*rhs._grid->_slice_stride[dimension];
@@ -131,7 +131,7 @@ template<class vobj> void Scatter_plane_simple (Lattice<vobj> &rhs,std::vector<v
int so = plane*rhs._grid->_ostride[dimension]; // base offset for start of plane
#pragma omp parallel for collapse(2)
PARALLEL_NESTED_LOOP(2)
for(int n=0;n<rhs._grid->_slice_nblock[dimension];n++){
for(int b=0;b<rhs._grid->_slice_block[dimension];b++){
@@ -160,7 +160,7 @@ template<class vobj> void Copy_plane(Lattice<vobj>& lhs,Lattice<vobj> &rhs, int
int ro = rplane*rhs._grid->_ostride[dimension]; // base offset for start of plane
int lo = lplane*lhs._grid->_ostride[dimension]; // base offset for start of plane
#pragma omp parallel for collapse(2)
PARALLEL_NESTED_LOOP(2)
for(int n=0;n<rhs._grid->_slice_nblock[dimension];n++){
for(int b=0;b<rhs._grid->_slice_block[dimension];b++){
@@ -185,7 +185,7 @@ template<class vobj> void Copy_plane_permute(Lattice<vobj>& lhs,Lattice<vobj> &r
int ro = rplane*rhs._grid->_ostride[dimension]; // base offset for start of plane
int lo = lplane*lhs._grid->_ostride[dimension]; // base offset for start of plane
#pragma omp parallel for collapse(2)
PARALLEL_NESTED_LOOP(2)
for(int n=0;n<rhs._grid->_slice_nblock[dimension];n++){
for(int b=0;b<rhs._grid->_slice_block[dimension];b++){
int o =n*rhs._grid->_slice_stride[dimension];