leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 29 Packages Projects Releases Wiki Activity

CUTLASS 2.4 (Implicit GEMM convolution) (#147)

Browse Source 
CUTLASS 2.4 (Implicit GEMM Convolution)

Co-authored-by: Manish Gupta <manigupta@nvidia.com>, Haicheng Wu <haichengw@nvidia.com>, Dustyn Blasig <dblasig@nvidia.com>, Andrew Kerr <akerr@nvidia.com>
This commit is contained in:
Manish Gupta
2020-11-19 21:25:25 -08:00
committed by GitHub
co-authored by Manish Gupta <manigupta@nvidia.com>, Haicheng Wu <haichengw@nvidia.com>, Dustyn Blasig <dblasig@nvidia.com>, Andrew Kerr <akerr@nvidia.com>
parent c2b80ad4e4
commit 6615010cd0
224 changed files with 43939 additions and 1061 deletions
Download Patch File Download Diff File Expand all files Collapse all files
examples/08_turing_tensorop_gemm/turing_tensorop_gemm.cu
+23 -29
View File
@@ -188,31 +188,6 @@ using Gemm = cutlass::gemm::device::Gemm<ElementInputA,
int run() {
// Turing Tensor Core operations exposed with mma.sync and ldmatrix are first available
// in CUDA 10.2.
//
// CUTLASS must be compiled with CUDA 10.2 Toolkit to run these examples.
if (!(__CUDACC_VER_MAJOR__ > 10 || (__CUDACC_VER_MAJOR__ == 10 && __CUDACC_VER_MINOR__ >= 2))) {
std::cerr << "Turing Tensor Core operations must be compiled with CUDA 10.2 Toolkit or later." << std::endl;
return -1;
}
cudaDeviceProp props;
cudaError_t error = cudaGetDeviceProperties(&props, 0);
if (error != cudaSuccess) {
std::cerr << "cudaGetDeviceProperties() returned an error: " << cudaGetErrorString(error) << std::endl;
return -1;
}
if (!((props.major * 10 + props.minor) >= 75)) {
std::cerr << "Turing Tensor Core operations must be run on a machine with compute capability at least 75."
<< std::endl;
// Return 0 so tests are considered passing if run on unsupported platforms.
return 0;
}
const int length_m = 5120;
const int length_n = 4096;
const int length_k = 4096;
@@ -337,18 +312,37 @@ int run() {
}
int main() {
bool notSupported = false;
// Turing Tensor Core operations exposed with mma.sync and ldmatrix are first available
// in CUDA 10.2.
//
// CUTLASS must be compiled with CUDA 10.2 Toolkit to run these examples.
if (!(__CUDACC_VER_MAJOR__ > 10 || (__CUDACC_VER_MAJOR__ == 10 && __CUDACC_VER_MINOR__ >= 2))) {
std::cerr << "Turing Tensor Core operations must be compiled with CUDA 10.2 Toolkit or later." << std::endl;
notSupported = true;
}
// Returning zero so this test passes when built on older Toolkits.
cudaDeviceProp props;
cudaError_t error = cudaGetDeviceProperties(&props, 0);
if (error != cudaSuccess) {
std::cerr << "cudaGetDeviceProperties() returned an error: " << cudaGetErrorString(error) << std::endl;
return -1;
}
if (!((props.major * 10 + props.minor) >= 75)) {
std::cerr << "Turing Tensor Core operations must be run on a machine with compute capability at least 75."
<< std::endl;
notSupported = true;
}
if (notSupported) {
// Returning zero so this test passes on older Toolkits. Its actions are no-op.
return 0;
}
else {
return run();
}
return run();
}