portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-09-20 01:05:38 +01:00
Code Releases Activity

sliceSumReduction_cub_large now also faster than CPU on Frontier

Browse Source
This commit is contained in:
Dennis Bollweg 2024-02-16 13:55:21 -05:00
parent e0d5e3c6c7
commit 15878f7613
2 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Grid/lattice/Lattice_slicesum_core.h
View File

@ -119,7 +119,7 @@ template<class vobj> inline void sliceSumReduction_cub_large(const vobj *Data, V
typedef typename vobj::vector_type vector;
const int words = sizeof(vobj)/sizeof(vector);
const int osites = rd*e1*e2;
Vector<vector>buffer(osites);
commVector<vector>buffer(osites);
vector *dat = (vector *)Data;
vector *buf = &buffer[0];
Vector<vector> lvSum_small(rd);

4
tests/core/Test_sliceSum.cc
View File

@ -105,6 +105,8 @@ int main (int argc, char ** argv) {
int trace_id = traceStart("sliceSum benchmark - ComplexD");
std::cout << GridLogMessage << "Testing ComplexD" << std::endl;
std::cout << GridLogMessage << "sizeof(ComplexD) = " << sizeof(ComplexD) << std::endl;
std::cout << GridLogMessage << "sizeof(vComplexD) = " << sizeof(vComplexD) << std::endl;
for (int i = 0; i < Nd; i++) {
RealD t=-usecond();
@ -258,4 +260,4 @@ int main (int argc, char ** argv) {
Grid_finalize();
return 0;
}
}