portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-09-20 09:15:38 +01:00
Code Releases Activity

Hit an annoying strict alias optimisation in GCC 4.9 through 6.3

Browse Source 
Chris K was correct. It appears that an additional memcpy (UGHHH) is enough
to suppress the compiler
This commit is contained in:
Peter Boyle 2018-03-07 07:27:26 -08:00
parent 40699221e2
commit 90a2efb9b3
1 changed files with 25 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
lib/tensors/Tensor_extract_merge.h
View File

@ -31,6 +31,10 @@ Author: Christopher Kelly <ckelly@phys.columbia.edu>
/* END LEGAL */
#pragma once
#include <string.h>
//#pragma GCC optimize("no-strict-aliasing")
NAMESPACE_BEGIN(Grid);
/////////////////////////////////////////////////////////////////
@ -44,7 +48,6 @@ template<class __T> using ExtractBuffer = AcceleratorVector<__T ,GRID_MAX
//void extract(const vobj &vec,ExtractPointerArray<sobj> &extracted, int offset);
//void merge(vobj &vec,ExtractBuffer<typename vobj::scalar_object> &extracted)
//void merge(vobj &vec,ExtractPointerArray<typename vobj::scalar_object> &extracted)
//
////////////////////////////////////////////////////////////////////////
// Extract to contiguous array scalar object
@ -60,12 +63,15 @@ void extract(const vobj &vec,ExtractBuffer<sobj> &extracted)
const int Nsimd=vector_type::Nsimd();
const int Nextr=extracted.size();
const int s=Nsimd/Nextr;
sobj_scalar_type *sp = (sobj_scalar_type *)&extracted[0];
sobj_scalar_type *sp = (sobj_scalar_type *) &extracted[0];
scalar_type *vp = (scalar_type *)&vec;
scalar_type vtmp;
sobj_scalar_type stmp;
for(int w=0;w<words;w++){
for(int i=0;i<Nextr;i++){
sp[i*words+w] = vp[w*Nsimd+i*s] ;
memcpy((char *)&vtmp,(char *)&vp[w*Nsimd+i*s],sizeof(vtmp));
stmp = vtmp;
memcpy((char *)&sp[i*words+w],(char *)&stmp,sizeof(stmp));
}
}
return;
@ -88,10 +94,14 @@ void merge(vobj &vec,ExtractBuffer<sobj> &extracted)
sobj_scalar_type *sp = (sobj_scalar_type *)&extracted[0];
scalar_type *vp = (scalar_type *)&vec;
scalar_type vtmp;
sobj_scalar_type stmp;
for(int w=0;w<words;w++){
for(int i=0;i<Nextr;i++){
for(int ii=0;ii<s;ii++){
vp[w*Nsimd+i*s+ii] = sp[i*words+w] ;
memcpy((char *)&stmp,(char *)&sp[i*words+w],sizeof(stmp));
vtmp = stmp;
memcpy((char *)&vp[w*Nsimd+i*s+ii],(char *)&vtmp,sizeof(vtmp));
}
}
}
@ -113,13 +123,14 @@ void extract(const vobj &vec,ExtractPointerArray<sobj> &extracted, int offset)
const int s = Nsimd/Nextr;
scalar_type * vp = (scalar_type *)&vec;
scalar_type vtmp;
sobj_scalar_type stmp;
for(int w=0;w<words;w++){
for(int i=0;i<Nextr;i++){
sobj_scalar_type * pointer = (sobj_scalar_type *)& extracted[i][offset];
scalar_type tmp =vp[w*Nsimd+i*s];
// assert(w*sizeof(sobj_scalar_type)<sizeof(sobj));
pointer[w] = tmp; // may do a precision conversion
memcpy((char *)&vtmp,(char *)&vp[w*Nsimd+i*s],sizeof(vtmp));
stmp = vtmp;
memcpy((char *)&pointer[w],(char *)&stmp,sizeof(stmp)); // may do a precision conversion
}
}
}
@ -140,11 +151,15 @@ void merge(vobj &vec,ExtractPointerArray<sobj> &extracted, int offset)
const int s = Nsimd/Nextr;
scalar_type * vp = (scalar_type *)&vec;
scalar_type vtmp;
sobj_scalar_type stmp;
for(int w=0;w<words;w++){
for(int i=0;i<Nextr;i++){
sobj_scalar_type * pointer = (sobj_scalar_type *)& extracted[i][offset];
for(int ii=0;ii<s;ii++){
vp[w*Nsimd+i*s+ii] = pointer[w] ;
memcpy((char *)&stmp,(char *)&pointer[w],sizeof(stmp));
vtmp=stmp;
memcpy((char *)&vp[w*Nsimd+i*s+ii],(char *)&vtmp,sizeof(vtmp));
}
}
}