eigen/unsupported/test/cxx11_tensor_contract_cuda.cu

217 lines
7.2 KiB
Plaintext
Raw Normal View History

2015-01-15 07:46:04 +08:00
// This file is part of Eigen, a lightweight C++ template library
// for linear algebra.
//
// Copyright (C) 2014 Benoit Steiner <benoit.steiner.goog@gmail.com>
// Copyright (C) 2014 Navdeep Jaitly <ndjaitly@google.com>
//
// This Source Code Form is subject to the terms of the Mozilla
// Public License v. 2.0. If a copy of the MPL was not distributed
// with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
#define EIGEN_TEST_NO_LONGDOUBLE
#define EIGEN_TEST_NO_COMPLEX
#define EIGEN_TEST_FUNC cxx11_tensor_cuda
#define EIGEN_DEFAULT_DENSE_INDEX_TYPE int
#define EIGEN_USE_GPU
#if defined __CUDACC_VER__ && __CUDACC_VER__ >= 70500
2016-09-01 19:39:33 +08:00
#include <cuda_fp16.h>
#endif
2015-01-15 07:46:04 +08:00
#include "main.h"
#include <unsupported/Eigen/CXX11/Tensor>
using Eigen::Tensor;
typedef Tensor<float, 1>::DimensionPair DimPair;
template<int DataLayout>
void test_cuda_contraction(int m_size, int k_size, int n_size)
2015-01-15 07:46:04 +08:00
{
2016-02-02 05:57:41 +08:00
std::cout << "Testing for (" << m_size << "," << k_size << "," << n_size << ")" << std::endl;
2015-01-15 07:46:04 +08:00
// with these dimensions, the output has 300 * 140 elements, which is
// more than 30 * 1024, which is the number of threads in blocks on
// a 15 SM GK110 GPU
2016-02-02 05:57:41 +08:00
Tensor<float, 2, DataLayout> t_left(m_size, k_size);
Tensor<float, 2, DataLayout> t_right(k_size, n_size);
Tensor<float, 2, DataLayout> t_result(m_size, n_size);
Tensor<float, 2, DataLayout> t_result_gpu(m_size, n_size);
2015-01-15 07:46:04 +08:00
Eigen::array<DimPair, 1> dims(DimPair(1, 0));
t_left.setRandom();
t_right.setRandom();
std::size_t t_left_bytes = t_left.size() * sizeof(float);
std::size_t t_right_bytes = t_right.size() * sizeof(float);
std::size_t t_result_bytes = t_result.size() * sizeof(float);
float* d_t_left;
float* d_t_right;
float* d_t_result;
cudaMalloc((void**)(&d_t_left), t_left_bytes);
cudaMalloc((void**)(&d_t_right), t_right_bytes);
cudaMalloc((void**)(&d_t_result), t_result_bytes);
cudaMemcpy(d_t_left, t_left.data(), t_left_bytes, cudaMemcpyHostToDevice);
cudaMemcpy(d_t_right, t_right.data(), t_right_bytes, cudaMemcpyHostToDevice);
Eigen::CudaStreamDevice stream;
2015-01-15 07:46:04 +08:00
Eigen::GpuDevice gpu_device(&stream);
Eigen::TensorMap<Eigen::Tensor<float, 2, DataLayout> >
gpu_t_left(d_t_left, Eigen::array<int, 2>(m_size, k_size));
Eigen::TensorMap<Eigen::Tensor<float, 2, DataLayout> >
gpu_t_right(d_t_right, Eigen::array<int, 2>(k_size, n_size));
Eigen::TensorMap<Eigen::Tensor<float, 2, DataLayout> >
gpu_t_result(d_t_result, Eigen::array<int, 2>(m_size, n_size));
gpu_t_result.device(gpu_device) = gpu_t_left.contract(gpu_t_right, dims);
t_result = t_left.contract(t_right, dims);
cudaMemcpy(t_result_gpu.data(), d_t_result, t_result_bytes, cudaMemcpyDeviceToHost);
for (size_t i = 0; i < t_result.size(); i++) {
if (fabs(t_result(i) - t_result_gpu(i)) < 1e-4f) {
continue;
2015-01-15 07:46:04 +08:00
}
if (Eigen::internal::isApprox(t_result(i), t_result_gpu(i), 1e-4f)) {
continue;
}
std::cout << "mismatch detected at index " << i << ": " << t_result(i)
<< " vs " << t_result_gpu(i) << std::endl;
assert(false);
2015-01-15 07:46:04 +08:00
}
cudaFree((void*)d_t_left);
cudaFree((void*)d_t_right);
cudaFree((void*)d_t_result);
}
template<int DataLayout>
void test_scalar(int m_size, int k_size, int n_size)
{
std::cout << "Testing for (" << m_size << "," << k_size << "," << n_size << ")" << std::endl;
// with these dimensions, the output has 300 * 140 elements, which is
// more than 30 * 1024, which is the number of threads in blocks on
// a 15 SM GK110 GPU
Tensor<float, 2, DataLayout> t_left(m_size, k_size);
Tensor<float, 2, DataLayout> t_right(k_size, n_size);
Tensor<float, 0, DataLayout> t_result;
Tensor<float, 0, DataLayout> t_result_gpu;
Eigen::array<DimPair, 2> dims(DimPair(0, 0), DimPair(1, 1));
t_left.setRandom();
t_right.setRandom();
std::size_t t_left_bytes = t_left.size() * sizeof(float);
std::size_t t_right_bytes = t_right.size() * sizeof(float);
std::size_t t_result_bytes = sizeof(float);
float* d_t_left;
float* d_t_right;
float* d_t_result;
cudaMalloc((void**)(&d_t_left), t_left_bytes);
cudaMalloc((void**)(&d_t_right), t_right_bytes);
cudaMalloc((void**)(&d_t_result), t_result_bytes);
cudaMemcpy(d_t_left, t_left.data(), t_left_bytes, cudaMemcpyHostToDevice);
cudaMemcpy(d_t_right, t_right.data(), t_right_bytes, cudaMemcpyHostToDevice);
Eigen::CudaStreamDevice stream;
Eigen::GpuDevice gpu_device(&stream);
Eigen::TensorMap<Eigen::Tensor<float, 2, DataLayout> >
gpu_t_left(d_t_left, m_size, k_size);
Eigen::TensorMap<Eigen::Tensor<float, 2, DataLayout> >
gpu_t_right(d_t_right, k_size, n_size);
Eigen::TensorMap<Eigen::Tensor<float, 0, DataLayout> >
gpu_t_result(d_t_result);
gpu_t_result.device(gpu_device) = gpu_t_left.contract(gpu_t_right, dims);
t_result = t_left.contract(t_right, dims);
cudaMemcpy(t_result_gpu.data(), d_t_result, t_result_bytes, cudaMemcpyDeviceToHost);
if (fabs(t_result() - t_result_gpu()) > 1e-4f &&
!Eigen::internal::isApprox(t_result(), t_result_gpu(), 1e-4f)) {
std::cout << "mismatch detected: " << t_result()
<< " vs " << t_result_gpu() << std::endl;
assert(false);
}
cudaFree((void*)d_t_left);
cudaFree((void*)d_t_right);
cudaFree((void*)d_t_result);
}
2016-02-02 05:57:41 +08:00
template<int DataLayout>
void test_cuda_contraction_m() {
2015-01-15 07:46:04 +08:00
for (int k = 32; k < 256; k++) {
2016-02-02 05:57:41 +08:00
test_cuda_contraction<ColMajor>(k, 128, 128);
test_cuda_contraction<RowMajor>(k, 128, 128);
2015-01-15 07:46:04 +08:00
}
2016-02-02 05:57:41 +08:00
}
template<int DataLayout>
void test_cuda_contraction_k() {
2015-01-15 07:46:04 +08:00
for (int k = 32; k < 256; k++) {
2016-02-02 05:57:41 +08:00
test_cuda_contraction<ColMajor>(128, k, 128);
test_cuda_contraction<RowMajor>(128, k, 128);
2015-01-15 07:46:04 +08:00
}
2016-02-02 05:57:41 +08:00
}
template<int DataLayout>
void test_cuda_contraction_n() {
2015-01-15 07:46:04 +08:00
for (int k = 32; k < 256; k++) {
2016-02-02 05:57:41 +08:00
test_cuda_contraction<ColMajor>(128, 128, k);
test_cuda_contraction<RowMajor>(128, 128, k);
2015-01-15 07:46:04 +08:00
}
2016-02-02 05:57:41 +08:00
}
2015-01-15 07:46:04 +08:00
2016-02-02 05:57:41 +08:00
template<int DataLayout>
void test_cuda_contraction_sizes() {
int m_sizes[] = { 31, 39, 63, 64, 65,
127, 129, 255, 257 , 511,
512, 513, 1023, 1024, 1025};
int n_sizes[] = { 31, 39, 63, 64, 65,
127, 129, 255, 257, 511,
512, 513, 1023, 1024, 1025};
2015-01-15 07:46:04 +08:00
2016-02-02 05:57:41 +08:00
int k_sizes[] = { 31, 39, 63, 64, 65,
95, 96, 127, 129, 255,
257, 511, 512, 513, 1023,
1024, 1025};
for (int i = 0; i < 15; i++) {
for (int j = 0; j < 15; j++) {
2015-01-15 07:46:04 +08:00
for (int k = 0; k < 17; k++) {
2016-02-02 05:57:41 +08:00
test_cuda_contraction<DataLayout>(m_sizes[i], n_sizes[j], k_sizes[k]);
2015-01-15 07:46:04 +08:00
}
}
}
2015-01-15 07:46:04 +08:00
}
2016-02-02 05:57:41 +08:00
void test_cxx11_tensor_cuda()
{
CALL_SUBTEST_1(test_cuda_contraction<ColMajor>(128, 128, 128));
CALL_SUBTEST_1(test_cuda_contraction<RowMajor>(128, 128, 128));
CALL_SUBTEST_1(test_scalar<ColMajor>(128, 128, 128));
CALL_SUBTEST_1(test_scalar<RowMajor>(128, 128, 128));
2016-02-02 05:57:41 +08:00
CALL_SUBTEST_2(test_cuda_contraction_m<ColMajor>());
CALL_SUBTEST_3(test_cuda_contraction_m<RowMajor>());
CALL_SUBTEST_4(test_cuda_contraction_k<ColMajor>());
CALL_SUBTEST_5(test_cuda_contraction_k<RowMajor>());
CALL_SUBTEST_6(test_cuda_contraction_n<ColMajor>());
CALL_SUBTEST_7(test_cuda_contraction_n<RowMajor>());
CALL_SUBTEST_8(test_cuda_contraction_sizes<ColMajor>());
CALL_SUBTEST_9(test_cuda_contraction_sizes<RowMajor>());
}