portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-19 08:17:05 +01:00
Code Releases Activity

fast MooeeInv for EOFA

Browse Source
This commit is contained in:
gfilaci
2019-09-02 14:26:13 +01:00
parent 0efaf3c4fa
commit e66669d300
2 changed files with 126 additions and 122 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
Grid/qcd/action/fermion/implementation/CayleyFermion5Dcache.h
View File

@ -148,7 +148,7 @@ CayleyFermion5D<Impl>::MooeeInv (const FermionField &psi_i, FermionField &chi
accelerator_for(sss,nloop,Simd::Nsimd(),{
uint64_t ss=sss*Ls;
typedef decltype(coalescedRead(psi[0])) spinor;
spinor tmp, acc, res;;
spinor tmp, acc, res;
// X = Nc*Ns
// flops = 2X + (Ls-2)(4X + 4X) + 6X + 1 + 2X + (Ls-1)(10X + 1) = -16X + Ls(1+18X) = -192 + 217*Ls flops