portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-12 20:27:06 +01:00
Code Releases Activity

Solves Hadrons compilation error adding extra templates

Browse Source
This commit is contained in:
TeseoSanJose
2025-02-11 09:48:07 +00:00
parent b50fb34e71
commit bdb302be66
1 changed files with 11 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
Grid/threads/Accelerator.h
View File

@ -215,7 +215,7 @@ inline void *acceleratorAllocHost(size_t bytes)
auto err = cudaMallocHost((void **)&ptr,bytes);
if( err != cudaSuccess ) {
ptr = (void *) NULL;
printf(" cudaMallocHost failed for %d %s \n",bytes,cudaGetErrorString(err));
printf(" cudaMallocHost failed for %ld %s \n",bytes,cudaGetErrorString(err));
assert(0);
}
return ptr;
@ -226,7 +226,7 @@ inline void *acceleratorAllocShared(size_t bytes)
auto err = cudaMallocManaged((void **)&ptr,bytes);
if( err != cudaSuccess ) {
ptr = (void *) NULL;
printf(" cudaMallocManaged failed for %d %s \n",bytes,cudaGetErrorString(err));
printf(" cudaMallocManaged failed for %ld %s \n",bytes,cudaGetErrorString(err));
assert(0);
}
return ptr;
@ -237,7 +237,7 @@ inline void *acceleratorAllocDevice(size_t bytes)
auto err = cudaMalloc((void **)&ptr,bytes);
if( err != cudaSuccess ) {
ptr = (void *) NULL;
printf(" cudaMalloc failed for %d %s \n",bytes,cudaGetErrorString(err));
printf(" cudaMalloc failed for %ld %s \n",bytes,cudaGetErrorString(err));
}
return ptr;
};
@ -245,7 +245,8 @@ inline void *acceleratorAllocDevice(size_t bytes)
inline void acceleratorFreeShared(void *ptr){ cudaFree(ptr);};
inline void acceleratorFreeDevice(void *ptr){ cudaFree(ptr);};
inline void acceleratorFreeHost(void *ptr){ cudaFree(ptr);};
inline void acceleratorCopyToDevice(void *from,void *to,size_t bytes) { cudaMemcpy(to,from,bytes, cudaMemcpyHostToDevice);}
inline void acceleratorCopyToDevice(void *from, void *to, size_t bytes) { cudaMemcpy(to,from,bytes, cudaMemcpyHostToDevice);}
inline void acceleratorCopyToDevice(const void *from, void *to, size_t bytes) {cudaMemcpy(to,from,bytes, cudaMemcpyHostToDevice);}
inline void acceleratorCopyFromDevice(void *from,void *to,size_t bytes){ cudaMemcpy(to,from,bytes, cudaMemcpyDeviceToHost);}
inline void acceleratorCopyToDeviceAsync(void *from, void *to, size_t bytes, cudaStream_t stream = copyStream) { cudaMemcpyAsync(to,from,bytes, cudaMemcpyHostToDevice, stream);}
inline void acceleratorCopyFromDeviceAsync(void *from, void *to, size_t bytes, cudaStream_t stream = copyStream) { cudaMemcpyAsync(to,from,bytes, cudaMemcpyDeviceToHost, stream);}
@ -649,7 +650,7 @@ accelerator_inline void acceleratorFence(void)
return;
}
inline void acceleratorCopyDeviceToDevice(void *from,void *to,size_t bytes)
inline void acceleratorCopyDeviceToDevice(void *from,void *to, size_t bytes)
{
acceleratorCopyDeviceToDeviceAsynch(from,to,bytes);
acceleratorCopySynchronise();
@ -657,7 +658,11 @@ inline void acceleratorCopyDeviceToDevice(void *from,void *to,size_t bytes)
template<class T> void acceleratorPut(T& dev,T&host)
{
acceleratorCopyToDevice(&host,&dev,sizeof(T));
acceleratorCopyToDevice(&host, &dev, sizeof(T));
}
template<class T> void acceleratorPut(T& dev, const T& host)
{
acceleratorCopyToDevice(&host, &dev, sizeof(T));
}
template<class T> T acceleratorGet(T& dev)
{