p47068915
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

add supports to op unpack on gpu

This commit is contained in:
zhouyuanshen 2020-11-04 15:40:25 +08:00
parent d79bcc923e
commit 5c4885821f
5 changed files with 398 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
mindspore/ccsrc/backend/kernel_compiler/gpu/arrays/unpack_gpu_kernel.cc Normal file
View File

@ -0,0 +1,40 @@
/**
* Copyright 2020 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "backend/kernel_compiler/gpu/arrays/unpack_gpu_kernel.h"
namespace mindspore {
namespace kernel {
MS_REG_GPU_KERNEL_ONE(
Unpack, KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeFloat32).AddOutputAttr(kNumberTypeFloat32),
UnpackGpuFwdKernel, float)
MS_REG_GPU_KERNEL_ONE(
Unpack, KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kNumberTypeFloat16),
UnpackGpuFwdKernel, half)
MS_REG_GPU_KERNEL_ONE(Unpack,
KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeInt32).AddOutputAttr(kNumberTypeInt32),
UnpackGpuFwdKernel, int)
MS_REG_GPU_KERNEL_ONE(Unpack,
KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeInt16).AddOutputAttr(kNumberTypeInt16),
UnpackGpuFwdKernel, int16_t)
MS_REG_GPU_KERNEL_ONE(Unpack,
KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeUInt8).AddOutputAttr(kNumberTypeUInt8),
UnpackGpuFwdKernel, uchar)
MS_REG_GPU_KERNEL_ONE(Unpack,
KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeBool).AddOutputAttr(kNumberTypeBool),
UnpackGpuFwdKernel, bool)
} // namespace kernel
} // namespace mindspore

112
mindspore/ccsrc/backend/kernel_compiler/gpu/arrays/unpack_gpu_kernel.h Normal file
View File

@ -0,0 +1,112 @@
/**
* Copyright 2020 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_GPU_UNPACK_GPU_KERNEL_H
#define MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_GPU_UNPACK_GPU_KERNEL_H
#include <vector>
#include <memory>
#include "backend/kernel_compiler/gpu/gpu_kernel.h"
#include "backend/kernel_compiler/gpu/gpu_kernel_factory.h"
#include "backend/kernel_compiler/gpu/cuda_impl/unpack.cuh"
namespace mindspore {
namespace kernel {
template <typename T>
class UnpackGpuFwdKernel : public GpuKernel {
public:
UnpackGpuFwdKernel() : axis_(0), output_num_(0), input_size_(1), dims_after_axis_(1), outputs_host_(nullptr) {}
~UnpackGpuFwdKernel() override = default;
const std::vector<size_t> &GetInputSizeList() const override { return input_size_list_; }
const std::vector<size_t> &GetOutputSizeList() const override { return output_size_list_; }
const std::vector<size_t> &GetWorkspaceSizeList() const override { return workspace_size_list_; }
bool Launch(const std::vector<AddressPtr> &inputs, const std::vector<AddressPtr> &workspace,
const std::vector<AddressPtr> &outputs, void *stream_ptr) override {
T *input = GetDeviceAddress<T>(inputs, 0);
T **outputs_array = GetDeviceAddress<T *>(workspace, 0);
for (size_t i = 0; i < outputs.size(); i++) {
outputs_host_[i] = GetDeviceAddress<T>(outputs, i);
}
CHECK_CUDA_RET_WITH_EXCEPT(cudaMemcpyAsync(outputs_array, outputs_host_.get(), sizeof(T *) * output_num_,
cudaMemcpyHostToDevice, reinterpret_cast<cudaStream_t>(stream_ptr)),
"Unpack opt cudaMemcpyAsync outputs failed");
UnpackKernel(SizeToInt(input_size_), output_num_, dims_after_axis_, outputs_array, input,
reinterpret_cast<cudaStream_t>(stream_ptr));
return true;
}
bool Init(const CNodePtr &kernel_node) override {
if (!CheckParam(kernel_node)) {
return false;
}
axis_ = static_cast<int32_t>(GetAttr<int64_t>(kernel_node, "axis"));
if (axis_ < 0) {
auto input_shape = AnfAlgo::GetInputDeviceShape(kernel_node, 0);
axis_ += SizeToInt(input_shape.size());
}
auto origin_data_format = AnfAlgo::GetOriginDataFormat(kernel_node);
auto input_format = AnfAlgo::GetInputFormat(kernel_node, 0);
axis_ = AxisTransform(origin_data_format, input_format, axis_);
output_num_ = static_cast<int32_t>(GetAttr<int64_t>(kernel_node, "num"));
outputs_host_ = std::make_unique<T *[]>(output_num_);
for (int i = 0; i < output_num_; i++) {
size_t _size = 1;
auto _shape = AnfAlgo::GetOutputDeviceShape(kernel_node, i);
for (size_t j = 0; j < _shape.size(); j++) {
_size *= _shape[j];
}
output_size_list_.push_back(_size * sizeof(T));
}
workspace_size_list_.push_back(sizeof(T *) * output_num_);
auto input_shape = AnfAlgo::GetInputDeviceShape(kernel_node, 0);
for (int i = 0; i < SizeToInt(input_shape.size()); i++) {
input_size_ *= input_shape[i];
if (i > axis_) {
dims_after_axis_ *= input_shape[i];
}
}
input_size_list_.push_back(input_size_ * sizeof(T));
InitSizeLists();
return true;
}
protected:
void InitSizeLists() override {}
private:
bool CheckParam(const CNodePtr &kernel_node) {
size_t input_num = AnfAlgo::GetInputTensorNum(kernel_node);
if (input_num != 1) {
MS_LOG(ERROR) << "input number is " << input_num << ", but UnpackGpuFwdKernel needs 1 input.";
return false;
}
return true;
}
int axis_;
int output_num_;
size_t input_size_;
int dims_after_axis_;
std::unique_ptr<T *[]> outputs_host_;
std::vector<size_t> input_size_list_;
std::vector<size_t> output_size_list_;
std::vector<size_t> workspace_size_list_;
};
} // namespace kernel
} // namespace mindspore
#endif // MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_GPU_UNPACK_GPU_KERNEL_H

59
mindspore/ccsrc/backend/kernel_compiler/gpu/cuda_impl/unpack.cu Executable file
View File

@ -0,0 +1,59 @@
/**
* Copyright 2020 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <stdio.h>
#include <stdint.h>
#include <cuda_runtime.h>
#include "backend/kernel_compiler/gpu/cuda_impl/unpack.cuh"
template <typename T>
__global__ void Unpack(const int size, const int output_num,
const int dims_after_axis, T** outputs, const T* input) {
for (int pos = blockIdx.x * blockDim.x + threadIdx.x; pos < (size); pos += blockDim.x * gridDim.x) {
int cycle = pos / (output_num * dims_after_axis);
int cur_output_index = pos % (output_num * dims_after_axis) / dims_after_axis;
int local_index = pos % (output_num * dims_after_axis) % dims_after_axis;
outputs[cur_output_index][cycle * dims_after_axis + local_index] = input[pos];
}
return;
}
template <typename T>
void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, T** outputs, const T* input,
cudaStream_t cuda_stream) {
Unpack<<<GET_BLOCKS(size), GET_THREADS, 0, cuda_stream>>>(size, output_num,
dims_after_axis, outputs, input);
return;
}
template void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, float** outputs, const float* input,
cudaStream_t cuda_stream);
template void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, half** outputs, const half* input,
cudaStream_t cuda_stream);
template void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, int** outputs, const int* input,
cudaStream_t cuda_stream);
template void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, int16_t** outputs, const int16_t* input,
cudaStream_t cuda_stream);
template void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, unsigned char** outputs, const unsigned char* input,
cudaStream_t cuda_stream);
template void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, bool** outputs, const bool* input,
cudaStream_t cuda_stream);

25
mindspore/ccsrc/backend/kernel_compiler/gpu/cuda_impl/unpack.cuh Executable file
View File

@ -0,0 +1,25 @@
/**
* Copyright 2020 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef MINDSPORE_CCSRC_KERNEL_GPU_CUDA_IMPL_UNPACKIMPL_H_
#define MINDSPORE_CCSRC_KERNEL_GPU_CUDA_IMPL_UNPACKIMPL_H_
#include "runtime/device/gpu/cuda_common.h"
template <typename T>
void UnpackKernel(const int size, const int output_num,
const int dims_after_axis, T** outputs, const T* input,
cudaStream_t cuda_stream);
#endif // MINDSPORE_CCSRC_KERNEL_GPU_CUDA_IMPL_UNPACKIMPL_H_

162
tests/st/ops/gpu/test_unpack_op.py Normal file
View File

@ -0,0 +1,162 @@
# Copyright 2020 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
import numpy as np
import pytest
import mindspore.context as context
import mindspore.nn as nn
import mindspore.ops.operations.array_ops as P
from mindspore import Tensor
from mindspore.common.api import ms_function
from mindspore.common.initializer import initializer
from mindspore.common.parameter import Parameter
class UnpackNet(nn.Cell):
def __init__(self, nptype):
super(UnpackNet, self).__init__()
self.unpack = P.Unpack(axis=3)
self.data_np = np.array([[[[[0, 0],
[0, 1]],
[[0, 0],
[2, 3]]],
[[[0, 0],
[4, 5]],
[[0, 0],
[6, 7]]]],
[[[[0, 0],
[8, 9]],
[[0, 0],
[10, 11]]],
[[[0, 0],
[12, 13]],
[[0, 0],
[14, 15]]]]]).astype(nptype)
self.x1 = Parameter(initializer(Tensor(self.data_np), [2, 2, 2, 2, 2]), name='x1')
@ms_function
def construct(self):
return self.unpack(self.x1)
def unpack(nptype):
context.set_context(mode=context.GRAPH_MODE, device_target='GPU')
unpack_ = UnpackNet(nptype)
output = unpack_()
expect = (np.reshape(np.array([0] * 16).astype(nptype), (2, 2, 2, 2)),
np.arange(2 * 2 * 2 * 2).reshape(2, 2, 2, 2).astype(nptype))
for i, exp in enumerate(expect):
assert (output[i].asnumpy() == exp).all()
def unpack_pynative(nptype):
context.set_context(mode=context.PYNATIVE_MODE, device_target='GPU')
x1 = np.array([[[[[0, 0],
[0, 1]],
[[0, 0],
[2, 3]]],
[[[0, 0],
[4, 5]],
[[0, 0],
[6, 7]]]],
[[[[0, 0],
[8, 9]],
[[0, 0],
[10, 11]]],
[[[0, 0],
[12, 13]],
[[0, 0],
[14, 15]]]]]).astype(nptype)
x1 = Tensor(x1)
expect = (np.reshape(np.array([0] * 16).astype(nptype), (2, 2, 2, 2)),
np.arange(2 * 2 * 2 * 2).reshape(2, 2, 2, 2).astype(nptype))
output = P.Unpack(axis=3)(x1)
for i, exp in enumerate(expect):
assert (output[i].asnumpy() == exp).all()
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_graph_float32():
unpack(np.float32)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_graph_float16():
unpack(np.float16)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_graph_int32():
unpack(np.int32)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_graph_int16():
unpack(np.int16)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_graph_uint8():
unpack(np.uint8)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_graph_bool():
unpack(np.bool)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_pynative_float32():
unpack_pynative(np.float32)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_pynative_float16():
unpack_pynative(np.float16)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_pynative_int32():
unpack_pynative(np.int32)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_pynative_int16():
unpack_pynative(np.int16)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_pynative_uint8():
unpack_pynative(np.uint8)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_unpack_pynative_bool():
unpack_pynative(np.bool)