leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 24 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
test/unit/gemm/threadblock/mma_multistage_sparse_testbed.h
+6 -8
View File
@@ -328,19 +328,17 @@ struct SparseTestbed {
test::gemm::threadblock::kernel_multistage_mma_sparse<Mma>,
cudaFuncAttributeMaxDynamicSharedMemorySize, smem_size);
EXPECT_EQ(result, cudaSuccess)
<< " cudaFuncSetAttribute "
"cudaFuncAttributeMaxDynamicSharedMemorySize error: "
<< cudaGetErrorString(result);
if (result != cudaSuccess) {
return true;
}
result = cudaFuncSetAttribute(
test::gemm::threadblock::kernel_multistage_mma_sparse<Mma>,
cudaFuncAttributePreferredSharedMemoryCarveout, 100);
EXPECT_EQ(result, cudaSuccess)
<< " cudaFuncSetAttribute "
"cudaFuncAttributePreferredSharedMemoryCarveout error: "
<< cudaGetErrorString(result);
if (result != cudaSuccess) {
return true;
}
}
test::gemm::threadblock::kernel_multistage_mma_sparse<Mma>
test/unit/gemm/threadblock/mma_multistage_testbed.h
+6 -8
View File
@@ -266,19 +266,17 @@ struct Testbed {
test::gemm::threadblock::kernel_multistage_mma<Mma>,
cudaFuncAttributeMaxDynamicSharedMemorySize, smem_size);
EXPECT_EQ(result, cudaSuccess)
<< " cudaFuncSetAttribute "
"cudaFuncAttributeMaxDynamicSharedMemorySize error: "
<< cudaGetErrorString(result);
if (result != cudaSuccess) {
return true;
}
result = cudaFuncSetAttribute(
test::gemm::threadblock::kernel_multistage_mma<Mma>,
cudaFuncAttributePreferredSharedMemoryCarveout, 100);
EXPECT_EQ(result, cudaSuccess)
<< " cudaFuncSetAttribute "
"cudaFuncAttributePreferredSharedMemoryCarveout error: "
<< cudaGetErrorString(result);
if (result != cudaSuccess) {
return true;
}
}
test::gemm::threadblock::kernel_multistage_mma<Mma>