portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-14 22:07:05 +01:00
Code Releases Activity

free malloc'd memory

Browse Source
This commit is contained in:
dbollweg
2024-02-22 21:42:44 -05:00
parent 6f3455900e
commit 1c8b807c2e
2 changed files with 10 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
Grid/lattice/Lattice_slicesum_core.h
View File

@ -5,6 +5,7 @@
#include <cub/cub.cuh>
#define gpucub cub
#define gpuMalloc cudaMalloc
#define gpuFree cudaFree
#define gpuMemcpyAsync cudaMemcpyAsync
#define gpuMemcpyDeviceToHost cudaMemcpyDeviceToHost
#define gpuMemcpyHostToDevice cudaMemcpyHostToDevice
@ -16,6 +17,7 @@
#include <hipcub/hipcub.hpp>
#define gpucub hipcub
#define gpuMalloc hipMalloc
#define gpuFree hipFree
#define gpuMemcpyAsync hipMemcpyAsync
#define gpuMemcpyDeviceToHost hipMemcpyDeviceToHost
#define gpuMemcpyHostToDevice hipMemcpyHostToDevice
@ -112,6 +114,10 @@ template<class vobj> inline void sliceSumReduction_cub_small(const vobj *Data, V
//sync after copy
accelerator_barrier();
gpuFree(temp_storage_array);
gpuFree(d_out);
gpuFree(d_offsets);
}
@ -201,7 +207,8 @@ template<class vobj> inline void sliceSumReduction_sycl(const Lattice<vobj> &Dat
theGridAccelerator->wait();
lvSum[r] = mysum[0];
}
free(mysum,*theGridAccelerator);
}
#endif