877bdcace6
CUTLASS 1.3 Release - Efficient GEMM kernel targeting Volta Tensor Cores via mma.sync instruction added in CUDA 10.1.
294 lines
9.8 KiB
Plaintext
294 lines
9.8 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 <cublas_v2.h>
|
|
#include <cstring>
|
|
#include "cutlass_unit_test.h"
|
|
|
|
#include "tools/util/half.h"
|
|
#include "tools/util/host_tensor.h"
|
|
#include "tools/util/tensor_view_io.h"
|
|
|
|
#include "cutlass/gemm/volta884_gemm_traits.h"
|
|
#include "cutlass/gemm/gemm.h"
|
|
|
|
#include "tools/test/unit/gemm/gemm_testbed.h"
|
|
#include "tools/test/unit/gemm/run_gemm.h"
|
|
|
|
#if CUTLASS_ENABLE_TENSOR_CORE_MMA
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x88x10_nn) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 88, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 16
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 88;
|
|
int partitionK_count = 10;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count);
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x88x10_nt) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 88, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 16
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 88;
|
|
int partitionK_count = 10;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count);
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x88x10_tn) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 88, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 16
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 88;
|
|
int partitionK_count = 10;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count);
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x88x10_tt) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 88, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 16
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 88;
|
|
int partitionK_count = 10;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count);
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x128x10_nn) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 128, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 12.
|
|
But if we require the partition mulitple to be 8, the first 9 partition
|
|
k = k - (k % partition_mulitiple) = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 56
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 128;
|
|
int partitionK_count = 10;
|
|
int partitionK_multiple = 8;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count, partitionK_multiple);
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x128x10_nt) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 128, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 12.
|
|
But if we require the partition mulitple to be 8, the first 9 partition
|
|
k = k - (k % partition_mulitiple) = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 56
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 128;
|
|
int partitionK_count = 10;
|
|
int partitionK_multiple = 8;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count, partitionK_multiple);
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x128x10_tn) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 128, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 12.
|
|
But if we require the partition mulitple to be 8, the first 9 partition
|
|
k = k - (k % partition_mulitiple) = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 56
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 128;
|
|
int partitionK_count = 10;
|
|
int partitionK_multiple = 8;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::MatrixLayout::kColumnMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count, partitionK_multiple);
|
|
}
|
|
|
|
////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
TEST(Volta884_h884gemm_partitionedK_64x64x32, volta884_h884gemm_128x256x128x10_tt) {
|
|
/*
|
|
for example
|
|
partitionedK gemm, m = 128, n = 256, overall_K = 128, partitionK_count = 10
|
|
for the first 9 partition k = overall_k / partitionK_count = 12.
|
|
But if we require the partition mulitple to be 8, the first 9 partition
|
|
k = k - (k % partition_mulitiple) = 8
|
|
for the last partition last_k = overall_k - (partitionK_count - 1) * k = 56
|
|
for volta884 it is safe to make sure leading dim are multiple of 8
|
|
*/
|
|
|
|
int m = 128;
|
|
int n = 256;
|
|
int overall_k = 128;
|
|
int partitionK_count = 10;
|
|
int partitionK_multiple = 8;
|
|
|
|
typedef cutlass::gemm::Volta884GemmTraits<
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::MatrixLayout::kRowMajor,
|
|
cutlass::Shape<32, 64, 64>,
|
|
cutlass::Shape<32, 64, 64>,
|
|
half,
|
|
half,
|
|
half,
|
|
2
|
|
> GemmTraits;
|
|
|
|
run_partitioned_k_gemm<GemmTraits>(m, n, overall_k, partitionK_count, partitionK_multiple);
|
|
}
|
|
|
|
#endif // if defined(CUTLASS_ENABLE_TENSOR_CORE_MMA)
|