portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-17 15:27:06 +01:00
Code Releases Activity

First compile against SYCL

Browse Source
This commit is contained in:
Peter Boyle
2020-05-05 11:13:27 -07:00
parent 04927d2e40
commit 28a1fcaaff
24 changed files with 205 additions and 87 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
Grid/tensors/Tensor_SIMT.h
View File

@ -34,14 +34,16 @@ NAMESPACE_BEGIN(Grid);
//accelerator_inline void SIMTsynchronise(void)
accelerator_inline void synchronise(void)
{
#ifdef __CUDA_ARCH__
#ifdef GRID_SIMT
#ifdef GRID_CUDA
// __syncthreads();
__syncwarp();
#endif
#endif
return;
}
#ifndef __CUDA_ARCH__
#ifndef GRID_SIMT
//////////////////////////////////////////
// Trivial mapping of vectors on host
//////////////////////////////////////////
@ -75,7 +77,13 @@ void coalescedWriteNonTemporal(vobj & __restrict__ vec,const vobj & __restrict__
vstream(vec, extracted);
}
#else
#ifdef GRID_CUDA
accelerator_inline int SIMTlane(int Nsimd) { return threadIdx.y; } // CUDA specific
#endif
#ifdef GRID_SYCL
//accelerator_inline int SIMTlane(int Nsimd) { return __spirv_BuiltInGlobalInvocationId[2]; } //SYCL specific
accelerator_inline int SIMTlane(int Nsimd) { return __spirv::initLocalInvocationId<3, cl::sycl::id<3>>()[2]; } // SYCL specific
#endif
//////////////////////////////////////////
// Extract and insert slices on the GPU

2
Grid/tensors/Tensor_exp.h
View File

@ -55,7 +55,7 @@ template<class vtype, int N> accelerator_inline iVector<vtype, N> Exponentiate(c
// Specialisation: Cayley-Hamilton exponential for SU(3)
#ifndef GRID_NVCC
#ifndef GRID_CUDA
template<class vtype, typename std::enable_if< GridTypeMapper<vtype>::TensorLevel == 0>::type * =nullptr>
accelerator_inline iMatrix<vtype,3> Exponentiate(const iMatrix<vtype,3> &arg, RealD alpha , Integer Nexp = DEFAULT_MAT_EXP )
{