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

Move coalesced read into tensors

Browse Source
This commit is contained in:
Peter Boyle 2019-05-25 12:43:00 +01:00
parent a0e9f3b0a0
commit aca788cf4f
1 changed files with 0 additions and 38 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
Grid/stencil/Stencil.h
View File

@ -56,44 +56,6 @@ NAMESPACE_BEGIN(Grid);
// Gather for when there *is* need to SIMD split with compression
///////////////////////////////////////////////////////////////////
#ifdef __CUDA_ARCH__
//////////////////////////////////////////
// EExtract and insert slices on the GPU
//////////////////////////////////////////
template<class vobj> accelerator_inline
typename vobj::scalar_object coalescedRead(const vobj & __restrict__ vec)
{
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_type vector_type;
constexpr int Nsimd = sizeof(vector_type)/sizeof(scalar_type);
int lane = threadIdx.x % Nsimd;
return extractLane(lane,vec);
}
template<class vobj> accelerator_inline
void coalescedWrite(vobj & __restrict__ vec,const typename vobj::scalar_object & __restrict__ extracted)
{
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_type vector_type;
constexpr int Nsimd = sizeof(vector_type)/sizeof(scalar_type);
int lane = threadIdx.x % Nsimd;
insertLane(lane,vec,extracted);
}
#else
//////////////////////////////////////////
// Trivial mapping of vectors on host
//////////////////////////////////////////
template<class vobj> accelerator_inline
vobj coalescedRead(const vobj & __restrict__ vec)
{
return vec;
}
template<class vobj> accelerator_inline
void coalescedWrite(vobj & __restrict__ vec,const vobj & __restrict__ extracted)
{
vec = extracted;
}
#endif
void Gather_plane_table_compute (GridBase *grid,int dimension,int plane,int cbmask,
int off,Vector<std::pair<int,int> > & table);