877bdcace6
CUTLASS 1.3 Release - Efficient GEMM kernel targeting Volta Tensor Cores via mma.sync instruction added in CUDA 10.1.
122 lines
4.7 KiB
Plaintext
122 lines
4.7 KiB
Plaintext
/***************************************************************************************************
|
|
* Copyright (c) 2017-2019, 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.
|
|
*
|
|
**************************************************************************************************/
|
|
|
|
#include "cutlass_unit_test.h"
|
|
#include "cutlass/gemm/gemm.h"
|
|
#include "cutlass/gemm/linear_scaling_device_ptr.h"
|
|
#include "cutlass/gemm/sgemm_traits.h"
|
|
|
|
#include "tools/test/unit/gemm/gemm_testbed.h"
|
|
#include "tools/test/unit/gemm/run_gemm.h"
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
// This example defines an SGEMM with a linear scaling functor that supports optionally passing
|
|
// alpha and beta via device-side pointers as in cuBLAS.
|
|
TEST(Sgemm_epilogue_functor, device_ptr_mode_sgemm_1024x512x128_nt) {
|
|
|
|
typedef cutlass::gemm::SgemmTraits<
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::Shape<8, 128, 128>,
|
|
cutlass::gemm::LinearScalingDevicePtr<float>
|
|
>
|
|
SgemmTraits;
|
|
|
|
// Define a GEMM problem size
|
|
int const m = 1025;
|
|
int const n = 512;
|
|
int const k = 128;
|
|
|
|
// Define scalars
|
|
float alpha_host = 3;
|
|
float beta_host = 2;
|
|
|
|
// Define a device-backed tensor to contain the scalars
|
|
cutlass::HostTensor<float, 1> device_scalars(2);
|
|
|
|
// Copy scalar values to device memory for device-ptr mode
|
|
device_scalars.at(0) = alpha_host;
|
|
device_scalars.at(1) = beta_host;
|
|
device_scalars.sync_device();
|
|
|
|
// Construct a GemmTestbed instance
|
|
test::GemmTestbed<
|
|
float, // AType
|
|
float, // BType
|
|
float, // CType
|
|
float, // Accumulator
|
|
float // Scalar
|
|
>
|
|
testbed(m,
|
|
n,
|
|
k,
|
|
test::convert(SgemmTraits::kLayoutA),
|
|
test::convert(SgemmTraits::kLayoutB),
|
|
alpha_host,
|
|
beta_host);
|
|
|
|
testbed.initialize();
|
|
|
|
//
|
|
// Construct a CUTLASS GEMM and initialize parameters
|
|
//
|
|
typedef cutlass::gemm::Gemm<SgemmTraits> Gemm;
|
|
typename Gemm::Params params;
|
|
|
|
params.initialize(testbed.M(),
|
|
testbed.N(),
|
|
testbed.K(),
|
|
0, // alpha ignored
|
|
testbed.ptr_A(),
|
|
testbed.lda(),
|
|
testbed.ptr_B(),
|
|
testbed.ldb(),
|
|
0, // beta ignored
|
|
testbed.ptr_C_initial(),
|
|
testbed.ldc(),
|
|
testbed.ptr_computed(),
|
|
testbed.ldc());
|
|
|
|
// Explicitly call the epilogue functor's initialize method to pass additional arguments
|
|
params.epilogue.functor.initialize(
|
|
device_scalars.device_data() + 0, // pointer to alpha in device memory
|
|
device_scalars.device_data() + 1); // pointer to beta in device memory
|
|
|
|
// Launch the CUTLASS SGEMM kernel
|
|
Gemm::launch(params);
|
|
|
|
// Report any errors
|
|
cudaError_t result = cudaDeviceSynchronize();
|
|
ASSERT_EQ(result, cudaSuccess)
|
|
<< "\nCUDA kernel launch error: " << cudaGetErrorString(result)
|
|
<< "\n";
|
|
|
|
// Verify result
|
|
ASSERT_TRUE(testbed.verify_with_cublas());
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|