portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-09-20 01:05:38 +01:00
Code Releases Activity

Accelerator memory

Browse Source
This commit is contained in:
Peter Boyle 2024-01-17 16:24:31 -05:00
parent a1f8bbb078
commit e07cb2b9de
1 changed files with 29 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
Grid/threads/Accelerator.h
View File

@ -117,7 +117,7 @@ accelerator_inline int acceleratorSIMTlane(int Nsimd) {
#endif #endif
} // CUDA specific } // CUDA specific
inline void cuda_mem(void) inline void acceleratorMem(void)
{ {
size_t free_t,total_t,used_t; size_t free_t,total_t,used_t;
cudaMemGetInfo(&free_t,&total_t); cudaMemGetInfo(&free_t,&total_t);
@ -125,6 +125,11 @@ inline void cuda_mem(void)
std::cout << " MemoryManager : GPU used "<<used_t<<" free "<<free_t<< " total "<<total_t<<std::endl; std::cout << " MemoryManager : GPU used "<<used_t<<" free "<<free_t<< " total "<<total_t<<std::endl;
} }
inline void cuda_mem(void)
{
acceleratorMem();
}
#define accelerator_for2dNB( iter1, num1, iter2, num2, nsimd, ... ) \ #define accelerator_for2dNB( iter1, num1, iter2, num2, nsimd, ... ) \
{ \ { \
int nt=acceleratorThreads(); \ int nt=acceleratorThreads(); \
@ -305,6 +310,11 @@ NAMESPACE_END(Grid);
NAMESPACE_BEGIN(Grid); NAMESPACE_BEGIN(Grid);
inline void acceleratorMem(void)
{
std::cout <<" SYCL acceleratorMem not implemented"<<std::endl;
}
extern cl::sycl::queue *theGridAccelerator; extern cl::sycl::queue *theGridAccelerator;
extern cl::sycl::queue *theCopyAccelerator; extern cl::sycl::queue *theCopyAccelerator;
@ -383,6 +393,15 @@ NAMESPACE_BEGIN(Grid);
#define accelerator __host__ __device__ #define accelerator __host__ __device__
#define accelerator_inline __host__ __device__ inline #define accelerator_inline __host__ __device__ inline
inline void acceleratorMem(void)
{
size_t free_t,total_t,used_t;
hipMemGetInfo(&free_t,&total_t);
used_t=total_t-free_t;
std::cout << " MemoryManager : GPU used "<<used_t<<" free "<<free_t<< " total "<<total_t<<std::endl;
}
extern hipStream_t copyStream; extern hipStream_t copyStream;
extern hipStream_t computeStream; extern hipStream_t computeStream;
/*These routines define mapping from thread grid to loop & vector lane indexing */ /*These routines define mapping from thread grid to loop & vector lane indexing */
@ -523,6 +542,15 @@ inline void acceleratorCopySynchronise(void) { auto discard=hipStreamSynchronize
#undef GRID_SIMT #undef GRID_SIMT
inline void acceleratorMem(void)
{
/*
struct rusage rusage;
getrusage( RUSAGE_SELF, &rusage );
return (size_t)rusage.ru_maxrss;
*/
std::cout <<" system acceleratorMem not implemented"<<std::endl;
}
#define accelerator #define accelerator
#define accelerator_inline strong_inline #define accelerator_inline strong_inline