portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-04-09 05:30:46 +01:00
Code Releases Activity

Accelerator loop changes for coalesced access

Browse Source
This commit is contained in:
Peter Boyle 2019-06-15 07:44:47 +01:00
parent cb93d32cd9
commit 81eb1fd9f2
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
Grid/algorithms/iterative/ConjugateGradient.h
View File

@ -132,9 +132,9 @@ public:
auto psi_v = psi.View(); auto psi_v = psi.View();
auto p_v = p.View(); auto p_v = p.View();
auto r_v = r.View(); auto r_v = r.View();
accelerator_loop(ss,p_v,{ accelerator_for(ss,p_v.size(), Field::vector_object::Nsimd(),{
vstream(psi_v[ss], a * p_v[ss] + psi_v[ss]); coalescedWrite(psi_v[ss], a * p_v(ss) + psi_v(ss));
vstream(p_v [ss], b * p_v[ss] + r_v[ss]); coalescedWrite(p_v[ss] , b * p_v(ss) + r_v (ss));
}); });
LinearCombTimer.Stop(); LinearCombTimer.Stop();
LinalgTimer.Stop(); LinalgTimer.Stop();