mirror of https://github.com/alibaba/MNN.git
164 lines
7.8 KiB
C++
164 lines
7.8 KiB
C++
//
|
|
// GridSampleBufExecution.cpp
|
|
// MNN
|
|
//
|
|
// Created by MNN on 2021/08/11.
|
|
// Copyright © 2018, Alibaba Group Holding Limited
|
|
//
|
|
|
|
#ifndef MNN_OPENCL_BUFFER_CLOSED
|
|
|
|
#include "backend/opencl/execution/buffer/GridSampleBufExecution.hpp"
|
|
|
|
namespace MNN {
|
|
namespace OpenCL {
|
|
GridSampleBufExecution::GridSampleBufExecution(const std::vector<Tensor *> &inputs, const MNN::Op *op, Backend *backend)
|
|
: CommonExecution(backend, op) {
|
|
mOpenCLBackend = static_cast<OpenCLBackend *>(backend);
|
|
mMode = op->main_as_GridSample()->mode();
|
|
mPaddingMode = op->main_as_GridSample()->paddingMode();
|
|
if (op->main_as_GridSample()->alignCorners()) {
|
|
mAlignCorners = 1;
|
|
}else {
|
|
mAlignCorners = 0;
|
|
}
|
|
}
|
|
|
|
ErrorCode GridSampleBufExecution::onEncode(const std::vector<Tensor *> &inputs, const std::vector<Tensor *> &outputs) {
|
|
auto runtime = mOpenCLBackend->getOpenCLRuntime();
|
|
auto inputTensor = inputs[0];
|
|
auto gridTensor = inputs[1];
|
|
auto outputTensor = outputs[0];
|
|
if(outputs[0]->dimensions() > 4){
|
|
mUnits.resize(1);
|
|
const int batches = inputTensor->buffer().dim[0].extent;
|
|
const int channels = inputTensor->buffer().dim[1].extent;
|
|
const int inD = inputTensor->buffer().dim[2].extent;
|
|
const int inH = inputTensor->buffer().dim[3].extent;
|
|
const int inW = inputTensor->buffer().dim[4].extent;
|
|
const int channelC4 = UP_DIV(channels, 4);
|
|
const int outD = outputTensor->buffer().dim[2].extent;
|
|
const int outH = outputTensor->buffer().dim[3].extent;
|
|
const int outW = outputTensor->buffer().dim[4].extent;
|
|
|
|
auto &unit = mUnits[0];
|
|
std::set<std::string> buildOptions;
|
|
if (mMode == SampleMode_BILINEAR) {
|
|
mKernelName = "bilinear5d_buf";
|
|
unit.kernel = runtime->buildKernel("grid_sample_buf", mKernelName, buildOptions, mOpenCLBackend->getPrecision());
|
|
} else {
|
|
mKernelName = "nearest5d_buf";
|
|
unit.kernel = runtime->buildKernel("grid_sample_buf", mKernelName, buildOptions, mOpenCLBackend->getPrecision());
|
|
}
|
|
mMaxWorkGroupSize = static_cast<uint32_t>(runtime->getMaxWorkGroupSize(unit.kernel));
|
|
mGlobalWorkSize = {
|
|
static_cast<uint32_t>(channelC4 * outD),
|
|
static_cast<uint32_t>(outW),
|
|
static_cast<uint32_t>(outH * batches)
|
|
};
|
|
MNN_ASSERT(outW > 0 && outH > 0);
|
|
|
|
uint32_t idx = 0;
|
|
cl_int ret = CL_SUCCESS;
|
|
ret |= unit.kernel->get().setArg(idx++, mGlobalWorkSize[0]);
|
|
ret |= unit.kernel->get().setArg(idx++, mGlobalWorkSize[1]);
|
|
ret |= unit.kernel->get().setArg(idx++, mGlobalWorkSize[2]);
|
|
ret |= unit.kernel->get().setArg(idx++, openCLBuffer(inputTensor));
|
|
ret |= unit.kernel->get().setArg(idx++, openCLBuffer(gridTensor));
|
|
ret |= unit.kernel->get().setArg(idx++, openCLBuffer(outputTensor));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(inH));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(inW));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(inD));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(outH));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(outW));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(outD));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(batches));
|
|
ret |= unit.kernel->get().setArg(idx++, mPaddingMode);
|
|
ret |= unit.kernel->get().setArg(idx++, mAlignCorners);
|
|
MNN_CHECK_CL_SUCCESS(ret, "setArg GridSampleBufExecution");
|
|
|
|
mLocalWorkSize = localWS3DDefault(mGlobalWorkSize, mMaxWorkGroupSize, runtime, mKernelName, unit.kernel, mOpenCLBackend->getCLTuneLevel(), "grid_sample_buf").first;
|
|
|
|
mOpenCLBackend->recordKernel3d(unit.kernel, mGlobalWorkSize, mLocalWorkSize);
|
|
unit.globalWorkSize = {mGlobalWorkSize[0], mGlobalWorkSize[1], mGlobalWorkSize[2]};
|
|
unit.localWorkSize = {mLocalWorkSize[0], mLocalWorkSize[1], mLocalWorkSize[2]};
|
|
}else{
|
|
mUnits.resize(1);
|
|
auto &unit = mUnits[0];
|
|
const int batches = inputTensor->buffer().dim[0].extent;
|
|
const int channels = inputTensor->buffer().dim[1].extent;
|
|
const int inH = inputTensor->buffer().dim[2].extent;
|
|
const int inW = inputTensor->buffer().dim[3].extent;
|
|
const int channelC4 = UP_DIV(channels, 4);
|
|
const int outH = outputTensor->buffer().dim[2].extent;
|
|
const int outW = outputTensor->buffer().dim[3].extent;
|
|
|
|
std::set<std::string> buildOptions;
|
|
if (mMode == 0) {
|
|
mKernelName = "bilinear_buf";
|
|
unit.kernel = runtime->buildKernel("grid_sample_buf", mKernelName, buildOptions, mOpenCLBackend->getPrecision());
|
|
}
|
|
else {
|
|
mKernelName = "nearest_buf";
|
|
unit.kernel = runtime->buildKernel("grid_sample_buf", mKernelName, buildOptions, mOpenCLBackend->getPrecision());
|
|
}
|
|
mMaxWorkGroupSize = static_cast<uint32_t>(runtime->getMaxWorkGroupSize(unit.kernel));
|
|
mGlobalWorkSize = {
|
|
static_cast<uint32_t>(channelC4),
|
|
static_cast<uint32_t>(outW),
|
|
static_cast<uint32_t>(outH * batches)
|
|
};
|
|
MNN_ASSERT(outW > 0 && outH > 0);
|
|
|
|
uint32_t idx = 0;
|
|
cl_int ret = CL_SUCCESS;
|
|
ret |= unit.kernel->get().setArg(idx++, mGlobalWorkSize[0]);
|
|
ret |= unit.kernel->get().setArg(idx++, mGlobalWorkSize[1]);
|
|
ret |= unit.kernel->get().setArg(idx++, mGlobalWorkSize[2]);
|
|
ret |= unit.kernel->get().setArg(idx++, openCLBuffer(inputTensor));
|
|
ret |= unit.kernel->get().setArg(idx++, openCLBuffer(gridTensor));
|
|
ret |= unit.kernel->get().setArg(idx++, openCLBuffer(outputTensor));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(inH));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(inW));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(outH));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(outW));
|
|
ret |= unit.kernel->get().setArg(idx++, static_cast<uint32_t>(batches));
|
|
ret |= unit.kernel->get().setArg(idx++, mPaddingMode);
|
|
ret |= unit.kernel->get().setArg(idx++, mAlignCorners);
|
|
MNN_CHECK_CL_SUCCESS(ret, "setArg GridSampleBufExecution");
|
|
|
|
mLocalWorkSize = localWS3DDefault(mGlobalWorkSize, mMaxWorkGroupSize, runtime, mKernelName, unit.kernel, mOpenCLBackend->getCLTuneLevel(), "grid_sample_buf").first;
|
|
|
|
mOpenCLBackend->recordKernel3d(unit.kernel, mGlobalWorkSize, mLocalWorkSize);
|
|
unit.globalWorkSize = {mGlobalWorkSize[0], mGlobalWorkSize[1], mGlobalWorkSize[2]};
|
|
unit.localWorkSize = {mLocalWorkSize[0], mLocalWorkSize[1], mLocalWorkSize[2]};
|
|
}
|
|
return NO_ERROR;
|
|
}
|
|
|
|
class GridSampleBufCreator :public OpenCLBackend::Creator {
|
|
public:
|
|
virtual ~GridSampleBufCreator() = default;
|
|
virtual Execution *onCreate(const std::vector<Tensor *> &inputs, const std::vector<Tensor *> &outputs,
|
|
const MNN::Op *op, Backend *backend) const override {
|
|
for (int i = 0; i < inputs.size(); ++i) {
|
|
TensorUtils::setTensorSupportPack(inputs[i], false);
|
|
}
|
|
for (int i = 0; i < outputs.size(); ++i) {
|
|
TensorUtils::setTensorSupportPack(outputs[i], false);
|
|
}
|
|
if (op->main_as_GridSample()->mode() != 0 && op->main_as_GridSample()->mode() != 1) {
|
|
MNN_PRINT("openCL buffer not support interpolate type: %d, fallback to cpu\n", op->main_as_GridSample()->mode());
|
|
return nullptr;
|
|
}
|
|
return new GridSampleBufExecution(inputs, op, backend);
|
|
}
|
|
};
|
|
|
|
REGISTER_OPENCL_OP_CREATOR(GridSampleBufCreator, OpType_GridSample, BUFFER);
|
|
|
|
} // namespace OpenCL
|
|
} // namespace MNN
|
|
|
|
#endif // MNN_OPENCL_BUFFER_CLOSED
|