mirror/CMake
1
0
Fork 0
mirror of https://github.com/Kitware/CMake.git synced 2026-02-21 14:40:26 -06:00
Code Issues Packages Projects Releases Wiki Activity

Tests/Cuda: Use memory allocated on the GPU in the kernels

Browse Source
This commit is contained in:
Pierre Moreau
2017-02-02 22:54:24 +01:00
committed by Brad King
parent 0ae5386aa9
commit 8731701cb6
2 changed files with 45 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
Tests/Cuda/Complex/file3.cu
View File

@@ -9,17 +9,25 @@ result_type_dynamic __device__ file2_func(int x);
static
__global__
void file3_kernel(result_type& r, int x)
void file3_kernel(result_type* r, int x)
{
r = file1_func(x);
*r = file1_func(x);
result_type_dynamic rd = file2_func(x);
}
int file3_launch_kernel(int x)
{
result_type r;
result_type* r;
cudaError_t err = cudaMallocManaged(&r, sizeof(result_type));
if(err != cudaSuccess)
{
std::cerr << "file3_launch_kernel: cudaMallocManaged failed: "
<< cudaGetErrorString(err) << std::endl;
return x;
}
file3_kernel <<<1,1>>> (r,x);
cudaError_t err = cudaGetLastError();
err = cudaGetLastError();
if(err != cudaSuccess)
{
std::cerr << "file3_kernel [SYNC] failed: "
@@ -33,5 +41,14 @@ int file3_launch_kernel(int x)
<< cudaGetErrorString(cudaGetLastError()) << std::endl;
return x;
}
return r.sum;
int result = r->sum;
err = cudaFree(r);
if(err != cudaSuccess)
{
std::cerr << "file3_launch_kernel: cudaFree failed: "
<< cudaGetErrorString(err) << std::endl;
return x;
}
return result;
}

28
Tests/Cuda/Complex/mixed.cu
View File

@@ -19,9 +19,9 @@ IMPORT void __host__ cuda_dynamic_lib_func();
static
__global__
void mixed_kernel(result_type& r, int x)
void mixed_kernel(result_type* r, int x)
{
r = file1_func(x);
*r = file1_func(x);
result_type_dynamic rd = file2_func(x);
}
@@ -29,9 +29,17 @@ EXPORT int mixed_launch_kernel(int x)
{
cuda_dynamic_lib_func();
result_type r;
result_type* r;
cudaError_t err = cudaMallocManaged(&r, sizeof(result_type));
if(err != cudaSuccess)
{
std::cerr << "mixed_launch_kernel: cudaMallocManaged failed: "
<< cudaGetErrorString(err) << std::endl;
return x;
}
mixed_kernel <<<1,1>>> (r,x);
cudaError_t err = cudaGetLastError();
err = cudaGetLastError();
if(err != cudaSuccess)
{
std::cerr << "mixed_kernel [SYNC] failed: "
@@ -45,5 +53,15 @@ EXPORT int mixed_launch_kernel(int x)
<< cudaGetErrorString(cudaGetLastError()) << std::endl;
return x;
}
return r.sum;
int result = r->sum;
err = cudaFree(r);
if(err != cudaSuccess)
{
std::cerr << "mixed_launch_kernel: cudaFree failed: "
<< cudaGetErrorString(err) << std::endl;
return x;
}
return result;
}