root
/
MNN
mirror of https://github.com/alibaba/MNN.git
1
0
Fork 0
Code Issues Actions 1 Packages Projects Releases Wiki Activity
MNN/source/backend/opencl/execution/buffer/RasterBufExecution.cpp

451 lines
18 KiB
C++
Raw Blame History

//
// RasterBufExecution.cpp
// MNN
//
// Created by MNN on 2020/05/12.
// Copyright © 2018, Alibaba Group Holding Limited
//
#ifndef MNN_OPENCL_BUFFER_CLOSED
#include "backend/opencl/execution/buffer/RasterBufExecution.hpp"
#include "core/Macro.h"
#include "core/TensorUtils.hpp"
#include "core/OpCommonUtils.hpp"
#include "backend/opencl/core/OpenCLBackend.hpp"
namespace MNN {
namespace OpenCL {
RasterBufExecution::RasterBufExecution(const std::vector<Tensor *> &inputs, const MNN::Op *op, Backend *backend)
: CommonExecution(backend, op) {
mOpenCLBackend = (OpenCLBackend *)backend;
//nothing to do
}
ErrorCode RasterBufExecution::onResize(const std::vector<Tensor *> &____inputs, const std::vector<Tensor *> &outputs) {
#ifdef LOG_VERBOSE
MNN_PRINT("start RasterBufExecution onResize !\n");
#endif
mTempInput.clear();
mTempOutput = nullptr;
MNN_ASSERT(outputs.size() == 1);
auto output = outputs[0];
if (!____inputs.empty()) {
OpCommonUtils::rasterInputReset(____inputs, outputs[0]);
}
auto des = TensorUtils::getDescribe(output);
auto outputDes = TensorUtils::getDescribe(output);
mNeedZero = !TensorUtils::regionIsFull(output);
auto regionNum = des->regions.size();
auto runtime = ((OpenCLBackend *)backend())->getOpenCLRuntime();
mFast = false;
if (outputDes->dimensionFormat == MNN_DATA_FORMAT_NC4HW4) {
mFast = true;
for (int i=0; i< des->regions.size(); ++i) {
auto& slice = des->regions[i];
if (TensorUtils::getDescribe(slice.origin)->dimensionFormat != MNN_DATA_FORMAT_NC4HW4) {
mFast = false;
break;
}
if (!OpCommonUtils::canBlitFast(slice, output)) {
mFast = false;
break;
}
}
}
if(mFast)
{
mUnits.resize(regionNum);
int kernel_idx = 0;
if(mNeedZero)
{
mUnits.resize(regionNum + 1);
auto outputShape = tensorShapeFormat(output);
int region[] = {outputShape[0], ROUND_UP(outputShape[3], 4), outputShape[1], outputShape[2]};//nc4hw
Unit &unit = mUnits[kernel_idx++];
unit.kernel = runtime->buildKernel("raster", "buffer_set_zero", {});
unit.localWorkSize = {8, 8};
unit.globalWorkSize = {(uint32_t)UP_DIV((region[2] * region[3]), 8)*8,
(uint32_t)UP_DIV((region[0] * region[1]), 8)*8};
int global_dim0 = region[2] * region[3];
int global_dim1 = region[0] * region[1];
uint32_t idx = 0;
cl_int ret = CL_SUCCESS;
ret |= unit.kernel.setArg(idx++, global_dim0);
ret |= unit.kernel.setArg(idx++, global_dim1);
ret |= unit.kernel.setArg(idx++, openCLBuffer(output));
if(ret != CL_SUCCESS)
{
MNN_PRINT("setArg err %d\n", (int)ret);
}
}
// nc4hw4 buffer raster
for (auto& slice : des->regions)
{
Tensor::InsideDescribe::Region C4Region;
OpCommonUtils::turnToPackRegion(slice, C4Region, output, 4);
Unit &unit = mUnits[kernel_idx++];
unit.kernel = runtime->buildKernel("raster_buf", "raster_nc4hw4_buffer", {});
const std::vector<uint32_t> gws = {(uint32_t)C4Region.size[2],
(uint32_t)C4Region.size[1],
(uint32_t)C4Region.size[0]};
uint32_t mMaxWorkGroupSize = static_cast<uint32_t>(runtime->getMaxWorkGroupSize(unit.kernel));
auto outputShape = tensorShapeFormat(output);
auto sliceShape = tensorShapeFormat(slice.origin);
uint32_t idx = 0;
cl_int ret = CL_SUCCESS;
ret |= unit.kernel.setArg(idx++, gws[0]);
ret |= unit.kernel.setArg(idx++, gws[1]);
ret |= unit.kernel.setArg(idx++, gws[2]);
ret |= unit.kernel.setArg(idx++, openCLBuffer(slice.origin));
ret |= unit.kernel.setArg(idx++, C4Region.src.offset);
ret |= unit.kernel.setArg(idx++, C4Region.src.stride[0]);
ret |= unit.kernel.setArg(idx++, C4Region.src.stride[1]);
ret |= unit.kernel.setArg(idx++, C4Region.src.stride[2]);
ret |= unit.kernel.setArg(idx++, sliceShape[1]);
ret |= unit.kernel.setArg(idx++, sliceShape[2]);
ret |= unit.kernel.setArg(idx++, sliceShape[3]);
ret |= unit.kernel.setArg(idx++, openCLBuffer(output));
ret |= unit.kernel.setArg(idx++, C4Region.dst.offset);
ret |= unit.kernel.setArg(idx++, C4Region.dst.stride[0]);
ret |= unit.kernel.setArg(idx++, C4Region.dst.stride[1]);
ret |= unit.kernel.setArg(idx++, C4Region.dst.stride[2]);
ret |= unit.kernel.setArg(idx++, outputShape[1]);
ret |= unit.kernel.setArg(idx++, outputShape[2]);
ret |= unit.kernel.setArg(idx++, outputShape[3]);
if(ret != CL_SUCCESS)
{
MNN_PRINT("setArg err %d\n", (int)ret);
}
std::string name = "raster_nc4hw4_buffer";
const std::vector<uint32_t> lws = localWS3DDefault(gws, mMaxWorkGroupSize, mOpenCLBackend->getOpenCLRuntime(), name, unit.kernel).first;
unit.localWorkSize = {lws[0], lws[1], lws[2]};
unit.globalWorkSize = {ROUND_UP(gws[0], std::max((uint32_t)1, lws[0])),
ROUND_UP(gws[1], std::max((uint32_t)1, lws[1])),
ROUND_UP(gws[2], std::max((uint32_t)1, lws[2]))};
}
if(mNeedZero)
{
MNN_ASSERT((regionNum+1==kernel_idx));
}
else
{
MNN_ASSERT((regionNum==kernel_idx));
}
return NO_ERROR;
}
bool cancombine = CanCombine(outputs);
// Alloc Temp buffer
auto bufferPool = ((OpenCLBackend *)backend())->getBufferPool();
auto bufferUnitSize = runtime->isSupportedFP16() ? sizeof(half_float::half) : sizeof(float);
for(int i=0; i< regionNum; ++i)
{
auto origin = des->regions[i].origin;
if(mTempInput.find(origin) != mTempInput.end())
{
continue;
}
auto buffer = bufferPool->alloc(origin->elementSize()*bufferUnitSize);
mTempInput.insert(std::make_pair(origin, buffer));
}
mTempOutput = bufferPool->alloc(output->elementSize() * bufferUnitSize);
for(auto& iter : mTempInput)
{
bufferPool->recycle(iter.second);
}
bufferPool->recycle(mTempOutput);
auto originNum = mTempInput.size();
if(cancombine){
regionNum = 1;
}
mUnits.resize(regionNum + originNum + 1);
int kernel_idx = 0;
if(mNeedZero)
{
mUnits.resize(regionNum + originNum + 2);
auto outputShape = tensorShapeFormat(output);
int region[] = {outputShape[0], ROUND_UP(outputShape[3], 4), outputShape[1], outputShape[2]};//nhwc
Unit &unit = mUnits[kernel_idx++];
unit.kernel = runtime->buildKernel("raster", "buffer_set_zero", {});
unit.localWorkSize = {8, 8};
unit.globalWorkSize = {(uint32_t)UP_DIV((region[2] * region[3]), 8)*8,
(uint32_t)UP_DIV((region[0] * region[1]), 8)*8};
int global_dim0 = region[2] * region[3];
int global_dim1 = region[0] * region[1];
uint32_t idx = 0;
cl_int ret = CL_SUCCESS;
ret |= unit.kernel.setArg(idx++, global_dim0);
ret |= unit.kernel.setArg(idx++, global_dim1);
ret |= unit.kernel.setArg(idx++, *mTempOutput);
if(ret != CL_SUCCESS)
{
MNN_PRINT("setArg err %d\n", (int)ret);
}
}
// nc4hw4 buffer to buffer
for(auto& iter : mTempInput)
{
Tensor* origin = iter.first;
std::vector<int> regionShape = tensorShapeFormat(origin);
int inputWH[] = {regionShape[2], regionShape[1]};
int region[] = {regionShape[0], UP_DIV(regionShape[3], 4), regionShape[1], regionShape[2]};
Unit &unit = mUnits[kernel_idx++];
if(TensorUtils::getDescribe(origin)->dimensionFormat == MNN_DATA_FORMAT_NHWC)// Image to nhwc buffer
{
unit.kernel = runtime->buildKernel("buffer_convert_buf", "nc4hw4_buffer_to_nhwc_buffer", {});
}
else //Image to nchw buffer
{
unit.kernel = runtime->buildKernel("buffer_convert_buf", "nc4hw4_buffer_to_nchw_buffer", {});
}
unit.localWorkSize = {16, 16};
unit.globalWorkSize = {(uint32_t)UP_DIV(region[3] * region[1], 16) * 16,
(uint32_t)UP_DIV(region[2] * region[0], 16) * 16};
int global_dim0 = region[3] * region[1];
int global_dim1 = region[2] * region[0];
//MNN_CHECK_CL_SUCCESS
uint32_t idx = 0;
cl_int ret = CL_SUCCESS;
ret |= unit.kernel.setArg(idx++, global_dim0);
ret |= unit.kernel.setArg(idx++, global_dim1);
ret |= unit.kernel.setArg(idx++, *(iter.second));
ret |= unit.kernel.setArg(idx++, inputWH[1]);
ret |= unit.kernel.setArg(idx++, inputWH[0]);
ret |= unit.kernel.setArg(idx++, regionShape[3]);
ret |= unit.kernel.setArg(idx++, openCLBuffer(origin));
if(ret != CL_SUCCESS)
{
MNN_PRINT("setArg err %d\n", (int)ret);
}
}
// buffer raster
if(cancombine){
auto regions = des->regions;
auto slice = regions[0];
int nums = regions.size();
int src_offset = regions[1].src.offset - slice.src.offset;
int dst_offset = regions[1].dst.offset - slice.dst.offset;
Unit &unit = mUnits[kernel_idx++];
unit.kernel = runtime->buildKernel("raster", "raster_buffer_combine", {});
unit.globalWorkSize = {(uint32_t)slice.size[2] * nums,
(uint32_t)slice.size[1],
(uint32_t)slice.size[0]};
const std::vector<uint32_t> gws = {(uint32_t)slice.size[2] * nums,
(uint32_t)slice.size[1],
(uint32_t)slice.size[0]};
uint32_t mMaxWorkGroupSize = static_cast<uint32_t>(runtime->getMaxWorkGroupSize(unit.kernel));
uint32_t idx = 0;
cl_int ret = CL_SUCCESS;
ret |= unit.kernel.setArg(idx++, gws[0]);
ret |= unit.kernel.setArg(idx++, gws[1]);
ret |= unit.kernel.setArg(idx++, gws[2]);
ret |= unit.kernel.setArg(idx++, *(mTempInput[slice.origin]));
ret |= unit.kernel.setArg(idx++, slice.src.offset);
ret |= unit.kernel.setArg(idx++, src_offset);
ret |= unit.kernel.setArg(idx++, slice.src.stride[0]);
ret |= unit.kernel.setArg(idx++, slice.src.stride[1]);
ret |= unit.kernel.setArg(idx++, slice.src.stride[2]);
ret |= unit.kernel.setArg(idx++, *mTempOutput);
ret |= unit.kernel.setArg(idx++, slice.dst.offset);
ret |= unit.kernel.setArg(idx++, dst_offset);
ret |= unit.kernel.setArg(idx++, slice.dst.stride[0]);
ret |= unit.kernel.setArg(idx++, slice.dst.stride[1]);
ret |= unit.kernel.setArg(idx++, slice.dst.stride[2]);
ret |= unit.kernel.setArg(idx++, slice.size[2]);
if(ret != CL_SUCCESS)
{
MNN_PRINT("setArg err %d\n", (int)ret);
}
std::string name = "rasterBuffer";
const std::vector<uint32_t> lws = localWS3DDefault(gws, mMaxWorkGroupSize, mOpenCLBackend->getOpenCLRuntime(), name, unit.kernel).first;
unit.localWorkSize = {lws[0], lws[1], lws[2]};
unit.globalWorkSize = {ROUND_UP(gws[0], std::max((uint32_t)1, lws[0])),
ROUND_UP(gws[1], std::max((uint32_t)1, lws[1])),
ROUND_UP(gws[2], std::max((uint32_t)1, lws[2]))};
recordKernel3d(unit.kernel, gws, lws, runtime);
}else{
for (auto& slice : des->regions)
{
Unit &unit = mUnits[kernel_idx++];
unit.kernel = runtime->buildKernel("raster_buf", "raster_buffer", {});
const std::vector<uint32_t> gws = {(uint32_t)slice.size[2],
(uint32_t)slice.size[1],
(uint32_t)slice.size[0]};
uint32_t mMaxWorkGroupSize = static_cast<uint32_t>(runtime->getMaxWorkGroupSize(unit.kernel));
uint32_t idx = 0;
cl_int ret = CL_SUCCESS;
ret |= unit.kernel.setArg(idx++, gws[0]);
ret |= unit.kernel.setArg(idx++, gws[1]);
ret |= unit.kernel.setArg(idx++, gws[2]);
ret |= unit.kernel.setArg(idx++, *(mTempInput[slice.origin]));
ret |= unit.kernel.setArg(idx++, slice.src.offset);
ret |= unit.kernel.setArg(idx++, slice.src.stride[0]);
ret |= unit.kernel.setArg(idx++, slice.src.stride[1]);
ret |= unit.kernel.setArg(idx++, slice.src.stride[2]);
ret |= unit.kernel.setArg(idx++, *mTempOutput);
ret |= unit.kernel.setArg(idx++, slice.dst.offset);
ret |= unit.kernel.setArg(idx++, slice.dst.stride[0]);
ret |= unit.kernel.setArg(idx++, slice.dst.stride[1]);
ret |= unit.kernel.setArg(idx++, slice.dst.stride[2]);
if(ret != CL_SUCCESS)
{
MNN_PRINT("setArg err %d\n", (int)ret);
}
std::string name = "raster_buffer";
const std::vector<uint32_t> lws = localWS3DDefault(gws, mMaxWorkGroupSize, mOpenCLBackend->getOpenCLRuntime(), name, unit.kernel).first;
unit.localWorkSize = {lws[0], lws[1], lws[2]};
unit.globalWorkSize = {ROUND_UP(gws[0], std::max((uint32_t)1, lws[0])),
ROUND_UP(gws[1], std::max((uint32_t)1, lws[1])),
ROUND_UP(gws[2], std::max((uint32_t)1, lws[2]))};
}
}
//buffer to nc4hw4 buffer
{
auto outputShape = tensorShapeFormat(output);
int wh[] = {outputShape[2], outputShape[1]};
int region[] = {outputShape[0], UP_DIV(outputShape[3], 4), outputShape[1], outputShape[2]};
Unit &unit = mUnits[kernel_idx++];
if(outputDes->dimensionFormat == MNN_DATA_FORMAT_NHWC)//nhwc buffer to Image
{
unit.kernel = runtime->buildKernel("buffer_convert_buf", "nhwc_buffer_to_nc4hw4_buffer", {});
}
else //nchw buffer to Image
{
unit.kernel = runtime->buildKernel("buffer_convert_buf", "nchw_buffer_to_nc4hw4_buffer", {});
}
unit.localWorkSize = {16, 16};
unit.globalWorkSize = {(uint32_t)UP_DIV(region[3] * region[1], 16) * 16,
(uint32_t)UP_DIV(region[2] * region[0], 16) * 16};
int global_dim0 = region[3] * region[1];
int global_dim1 = region[2] * region[0];
uint32_t idx = 0;
cl_int ret = CL_SUCCESS;
ret |= unit.kernel.setArg(idx++, global_dim0);
ret |= unit.kernel.setArg(idx++, global_dim1);
ret |= unit.kernel.setArg(idx++, *mTempOutput);
ret |= unit.kernel.setArg(idx++, wh[1]);
ret |= unit.kernel.setArg(idx++, wh[0]);
ret |= unit.kernel.setArg(idx++, outputShape[3]);
ret |= unit.kernel.setArg(idx++, openCLBuffer(output));
if(ret != CL_SUCCESS)
{
MNN_PRINT("setArg err %d\n", (int)ret);
}
}
//kernel num check
if(mNeedZero)
{
MNN_ASSERT((kernel_idx==regionNum + originNum + 2));
}
else
{
MNN_ASSERT((kernel_idx==regionNum + originNum + 1));
}
#ifdef LOG_VERBOSE
MNN_PRINT("end RasterBufExecution onResize !\n");
#endif
return NO_ERROR;
}
class RasterCreator : public OpenCLBackend::Creator {
public:
virtual ~RasterCreator() = 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);
}
return new RasterBufExecution(inputs, op, backend);
}
};
bool RasterBufExecution::CanCombine(const std::vector<Tensor *> &outputs){
auto des = TensorUtils::getDescribe(outputs[0]);
auto regions = des->regions;
if(regions.size() < 2)
return false;
auto origin = regions[0].origin;
const int size0 = regions[0].size[0];
const int size1 = regions[0].size[1];
const int size2 = regions[0].size[2];
const int src_offset = regions[1].src.offset - regions[0].src.offset;
const int dst_offset = regions[1].dst.offset - regions[0].dst.offset;
const int src_sride0 = regions[0].src.stride[0];
const int src_sride1 = regions[0].src.stride[1];
const int src_sride2 = regions[0].src.stride[2];
const int dst_sride0 = regions[0].dst.stride[0];
const int dst_sride1 = regions[0].dst.stride[1];
const int dst_sride2 = regions[0].dst.stride[2];
bool res = true;
for(int i = 1; i < regions.size(); ++i){
res &= regions[i].origin == origin;
res &= regions[i].size[0] == size0;
res &= regions[i].size[1] == size1;
res &= regions[i].size[2] == size2;
res &= regions[i].src.stride[0] == src_sride0;
res &= regions[i].src.stride[1] == src_sride1;
res &= regions[i].src.stride[2] == src_sride2;
res &= regions[i].dst.stride[0] == dst_sride0;
res &= regions[i].dst.stride[1] == dst_sride1;
res &= regions[i].dst.stride[2] == dst_sride2;
res &= (regions[i].src.offset - regions[i - 1].src.offset) == src_offset;
res &= (regions[i].dst.offset - regions[i - 1].dst.offset) == dst_offset;
if(res == false){
return res;
}
}
return res;
}
OpenCLCreatorRegister<RasterCreator> __RasterBuf_op(OpType_Raster, BUFFER);
} // namespace OpenCL
} // namespace MNN
#endif /* MNN_OPENCL_BUFFER_CLOSED */
Reference in New Issue View Git Blame Copy Permalink