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

Make arch tag of test cases more precisely to SM60

Browse Source 
Signed-off-by: Peter Han <fujun.han@iluvatar.ai>
This commit is contained in:
Peter Han 2021-03-05 10:53:26 +08:00
parent a3639ab1a0
commit 6c4539e372
2 changed files with 12 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
test/unit/conv/device/CMakeLists.txt
View File

@ -103,7 +103,7 @@ cutlass_test_unit_add_executable(
conv2d_wgrad_implicit_gemm_cf32nhwc_cf32nhwc_cf32nhwc_simt_f32_sm50.cu
# F16
conv2d_fprop_implicit_gemm_f16nhwc_f16nhwc_f16nhwc_simt_f16_sm50.cu
conv2d_fprop_implicit_gemm_f16nhwc_f16nhwc_f16nhwc_simt_f16_sm60.cu
)
if (CUTLASS_NVCC_MAX_ARCH GREATER_EQUAL 80)

22
test/unit/conv/device/conv2d_fprop_implicit_gemm_f16nhwc_f16nhwc_f16nhwc_simt_f16_sm50.cu → test/unit/conv/device/conv2d_fprop_implicit_gemm_f16nhwc_f16nhwc_f16nhwc_simt_f16_sm60.cu
View File

@ -37,7 +37,7 @@
////////////////////////////////////////////////////////////////////////////////
TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt_f16,
TEST(SM60_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt_f16,
32x64_8x2_32x32x8) {
/// Conv operation element types for the Gemm equivalent (ImplicitGemm)
@ -50,15 +50,15 @@ TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt
/// Device-level Conv2d instance
using Conv2dFpropKernel = typename cutlass::conv::kernel::DefaultConv2dFprop<
ElementA,
ElementA,
cutlass::layout::TensorNHWC,
ElementB,
ElementB,
cutlass::layout::TensorNHWC,
ElementC,
ElementC,
cutlass::layout::TensorNHWC,
ElementAccumulator,
cutlass::arch::OpClassSimt,
cutlass::arch::Sm50,
cutlass::arch::Sm60,
cutlass::gemm::GemmShape<32, 64, 8>,
cutlass::gemm::GemmShape<32, 32, 8>,
cutlass::gemm::GemmShape<1, 1, 1>,
@ -83,7 +83,7 @@ TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt
////////////////////////////////////////////////////////////////////////////////
TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt_f16,
TEST(SM60_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt_f16,
32x128_8x2_16x64x8) {
/// Conv operation element types for the Gemm equivalent (ImplicitGemm)
@ -96,15 +96,15 @@ TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt
/// Device-level Conv2d instance
using Conv2dFpropKernel = typename cutlass::conv::kernel::DefaultConv2dFprop<
ElementA,
ElementA,
cutlass::layout::TensorNHWC,
ElementB,
ElementB,
cutlass::layout::TensorNHWC,
ElementC,
ElementC,
cutlass::layout::TensorNHWC,
ElementAccumulator,
cutlass::arch::OpClassSimt,
cutlass::arch::Sm50,
cutlass::arch::Sm60,
cutlass::gemm::GemmShape<32, 128, 8>,
cutlass::gemm::GemmShape<16, 64, 8>,
cutlass::gemm::GemmShape<1, 1, 1>,
@ -116,7 +116,7 @@ TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt
>,
cutlass::gemm::threadblock::GemmIdentityThreadblockSwizzle<4>,
2,
cutlass::arch::OpMultiplyAddComplex,
cutlass::arch::OpMultiplyAdd,
cutlass::conv::IteratorAlgorithm::kOptimized
>::Kernel;