portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-10 07:55:35 +00:00
Code Releases Activity

HIP fix

Browse Source
This commit is contained in:
Peter Boyle 2020-11-01 01:16:01 +00:00
parent f313565a3c
commit 5eeabaa2bb
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
Grid/lattice/Lattice_basis.h
View File

@ -62,7 +62,7 @@ void basisRotate(VField &basis,Matrix& Qt,int j0, int j1, int k0,int k1,int Nm)
basis_v.push_back(basis[k].View(AcceleratorWrite)); basis_v.push_back(basis[k].View(AcceleratorWrite));
} }
#if ( (!defined(GRID_SYCL)) && (!defined(GRID_CUDA)) && (!defined(GRID_HIP)) ) #if ( (!defined(GRID_SYCL)) && (!defined(GRID_CUDA)) )
int max_threads = thread_max(); int max_threads = thread_max();
Vector < vobj > Bt(Nm * max_threads); Vector < vobj > Bt(Nm * max_threads);
thread_region thread_region
@ -161,11 +161,12 @@ void basisRotateJ(Field &result,std::vector<Field> &basis,Eigen::MatrixXd& Qt,in
double * Qt_j = & Qt_jv[0]; double * Qt_j = & Qt_jv[0];
for(int k=0;k<Nm;++k) Qt_j[k]=Qt(j,k); for(int k=0;k<Nm;++k) Qt_j[k]=Qt(j,k);
auto basis_vp=& basis_v[0];
autoView(result_v,result,AcceleratorWrite); autoView(result_v,result,AcceleratorWrite);
accelerator_for(ss, grid->oSites(),vobj::Nsimd(),{ accelerator_for(ss, grid->oSites(),vobj::Nsimd(),{
auto B=coalescedRead(zz); auto B=coalescedRead(zz);
for(int k=k0; k<k1; ++k){ for(int k=k0; k<k1; ++k){
B +=Qt_j[k] * coalescedRead(basis_v[k][ss]); B +=Qt_j[k] * coalescedRead(basis_vp[k][ss]);
} }
coalescedWrite(result_v[ss], B); coalescedWrite(result_v[ss], B);
}); });