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

Append fp16 test case to verify Mma_HFMA2

Browse Source 
Signed-off-by: Peter Han <fujun.han@iluvatar.ai>
This commit is contained in:
Peter Han 2021-03-04 18:17:57 +08:00
parent 169181f30f
commit a3639ab1a0
2 changed files with 131 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
test/unit/conv/device/CMakeLists.txt
View File

@ -101,6 +101,9 @@ cutlass_test_unit_add_executable(
conv2d_fprop_implicit_gemm_cf32nhwc_cf32nhwc_cf32nhwc_simt_f32_sm50.cu
conv2d_dgrad_implicit_gemm_cf32nhwc_cf32nhwc_cf32nhwc_simt_f32_sm50.cu
conv2d_wgrad_implicit_gemm_cf32nhwc_cf32nhwc_cf32nhwc_simt_f32_sm50.cu
# F16
conv2d_fprop_implicit_gemm_f16nhwc_f16nhwc_f16nhwc_simt_f16_sm50.cu
)
if (CUTLASS_NVCC_MAX_ARCH GREATER_EQUAL 80)

128
test/unit/conv/device/conv2d_fprop_implicit_gemm_f16nhwc_f16nhwc_f16nhwc_simt_f16_sm50.cu Normal file
View File

@ -0,0 +1,128 @@
/***************************************************************************************************
* Copyright (c) 2017-2021, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without modification, are permitted
* provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright notice, this list of
* conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright notice, this list of
* conditions and the following disclaimer in the documentation and/or other materials
* provided with the distribution.
* * Neither the name of the NVIDIA CORPORATION nor the names of its contributors may be used
* to endorse or promote products derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
* FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
* OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TOR (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
**************************************************************************************************/
/*! \file
\brief Tests for device-wide Implicit GEMM interface
*/
#include "../../common/cutlass_unit_test.h"
#include "cutlass/cutlass.h"
#include "cutlass/conv/kernel/default_conv2d_fprop.h"
#include "cutlass/conv/device/implicit_gemm_convolution.h"
#include "conv2d_testbed.h"
////////////////////////////////////////////////////////////////////////////////
TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt_f16,
32x64_8x2_32x32x8) {
/// Conv operation element types for the Gemm equivalent (ImplicitGemm)
using ElementA = cutlass::half_t;
using ElementB = cutlass::half_t;
using ElementC = cutlass::half_t;
using ElementAccumulator = cutlass::half_t;
using ElementCompute = cutlass::half_t;
/// Device-level Conv2d instance
using Conv2dFpropKernel = typename cutlass::conv::kernel::DefaultConv2dFprop<
ElementA,
cutlass::layout::TensorNHWC,
ElementB,
cutlass::layout::TensorNHWC,
ElementC,
cutlass::layout::TensorNHWC,
ElementAccumulator,
cutlass::arch::OpClassSimt,
cutlass::arch::Sm50,
cutlass::gemm::GemmShape<32, 64, 8>,
cutlass::gemm::GemmShape<32, 32, 8>,
cutlass::gemm::GemmShape<1, 1, 1>,
cutlass::epilogue::thread::LinearCombination<
ElementC,
1,
ElementAccumulator,
ElementCompute
>,
cutlass::gemm::threadblock::GemmIdentityThreadblockSwizzle<>,
2,
cutlass::arch::OpMultiplyAdd,
cutlass::conv::IteratorAlgorithm::kAnalytic
>::Kernel;
using Conv2dFprop = cutlass::conv::device::ImplicitGemmConvolution<Conv2dFpropKernel>;
/// Run all unit test sizes with device-level Conv2d instance
EXPECT_TRUE(test::conv::device::TestAllConv2d<Conv2dFprop>());
}
////////////////////////////////////////////////////////////////////////////////
TEST(SM50_Device_Conv2d_Fprop_Analytic_ImplicitGemm_f16nhwc_f16nhwc_f16nhwc_simt_f16,
32x128_8x2_16x64x8) {
/// Conv operation element types for the Gemm equivalent (ImplicitGemm)
using ElementA = cutlass::half_t;
using ElementB = cutlass::half_t;
using ElementC = cutlass::half_t;
using ElementAccumulator = cutlass::half_t;
using ElementCompute = cutlass::half_t;
/// Device-level Conv2d instance
using Conv2dFpropKernel = typename cutlass::conv::kernel::DefaultConv2dFprop<
ElementA,
cutlass::layout::TensorNHWC,
ElementB,
cutlass::layout::TensorNHWC,
ElementC,
cutlass::layout::TensorNHWC,
ElementAccumulator,
cutlass::arch::OpClassSimt,
cutlass::arch::Sm50,
cutlass::gemm::GemmShape<32, 128, 8>,
cutlass::gemm::GemmShape<16, 64, 8>,
cutlass::gemm::GemmShape<1, 1, 1>,
cutlass::epilogue::thread::LinearCombination<
ElementC,
1,
ElementAccumulator,
ElementCompute
>,
cutlass::gemm::threadblock::GemmIdentityThreadblockSwizzle<4>,
2,
cutlass::arch::OpMultiplyAddComplex,
cutlass::conv::IteratorAlgorithm::kOptimized
>::Kernel;
using Conv2dFprop = cutlass::conv::device::ImplicitGemmConvolution<Conv2dFpropKernel>;
/// Run all unit test sizes with device-level Conv2d instance
EXPECT_TRUE(test::conv::device::TestAllConv2d<Conv2dFprop>());
}