portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-04-09 21:50:45 +01:00
Code Releases Activity

Bug fix with spreadout FFT

Browse Source
This commit is contained in:
paboyle 2017-09-21 11:10:08 +01:00
parent a9ec5cf564
commit fddeb29d6b
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
lib/algorithms/FFT.h
View File

@ -230,6 +230,7 @@ namespace Grid {
// Barrel shift and collect global pencil // Barrel shift and collect global pencil
std::vector<int> lcoor(Nd), gcoor(Nd); std::vector<int> lcoor(Nd), gcoor(Nd);
result = source; result = source;
int pc = processor_coor[dim];
for(int p=0;p<processors[dim];p++) { for(int p=0;p<processors[dim];p++) {
PARALLEL_REGION PARALLEL_REGION
{ {
@ -240,7 +241,8 @@ namespace Grid {
for(int idx=0;idx<sgrid->lSites();idx++) { for(int idx=0;idx<sgrid->lSites();idx++) {
sgrid->LocalIndexToLocalCoor(idx,cbuf); sgrid->LocalIndexToLocalCoor(idx,cbuf);
peekLocalSite(s,result,cbuf); peekLocalSite(s,result,cbuf);
cbuf[dim]+=p*L; cbuf[dim]+=((pc+p) % processors[dim])*L;
// cbuf[dim]+=p*L;
pokeLocalSite(s,pgbuf,cbuf); pokeLocalSite(s,pgbuf,cbuf);
} }
} }
@ -278,7 +280,6 @@ namespace Grid {
flops+= flops_call*NN; flops+= flops_call*NN;
// writing out result // writing out result
int pc = processor_coor[dim];
PARALLEL_REGION PARALLEL_REGION
{ {
std::vector<int> clbuf(Nd), cgbuf(Nd); std::vector<int> clbuf(Nd), cgbuf(Nd);