Skip to content

P
Paddle-Lite

PaddlePaddle / Paddle-Lite

通知 331
Star 4
Fork 1
P
Paddle-Lite
提交 99ee4ff3 编写于 作者: 开心的小妮's avatar 开心的小妮
浏览文件
操作

add left code. test=develop

上级 a59c5c10
隐藏空白更改
内联 并排
Showing with 58 addition and 52 deletion
lite/backends/opencl/cl_runtime.cc
浏览文件 @ 99ee4ff3
......@@ -49,6 +49,8 @@ void CLRuntime::ReleaseResources() {
command_queue_->flush();
command_queue_->finish();
}
LOG(INFO) << "kernels_.size():" << kernels_.size();
LOG(INFO) << "programs_.size():" << programs_.size();
for (size_t kidx = 0; kidx < kernels_.size(); ++kidx) {
clReleaseKernel(kernels_[kidx]->get());
kernels_[kidx].reset();
......@@ -56,9 +58,9 @@ void CLRuntime::ReleaseResources() {
kernels_.clear();
kernel_offset_.clear();
for (auto& p : programs_) {
clReleaseProgram(p.second->get());
// clReleaseProgram(p.second->get());
}
programs_.clear();
// programs_.clear();
LOG(INFO) << "release resources finished.";
is_resources_released_ = true;
}
......
lite/kernels/opencl/layout_image_compute.cc
浏览文件 @ 99ee4ff3
......@@ -37,6 +37,11 @@ class LayoutComputeBufferChwToImageDefault
public:
using param_t = operators::LayoutParam;
~LayoutComputeBufferChwToImageDefault() {
LOG(INFO) << "Release LayoutComputeBufferChwToImageDefault";
kernel_.reset();
event_.reset();
}
void PrepareForRun() override {
auto& param = Param<param_t>();
if (param.process_type == 1) {
......@@ -44,7 +49,7 @@ class LayoutComputeBufferChwToImageDefault
}
VLOG(1) << "kernel_func_name_:" << kernel_func_name_;
auto& context = ctx_->As<OpenCLContext>();
context.cl_context()->AddKernel(
kernel_ = context.cl_context()->CreateKernel(
kernel_func_name_, "image/layout_kernel.cl", build_options_);
}
......@@ -90,40 +95,37 @@ class LayoutComputeBufferChwToImageDefault
VLOG(2) << "Stride2:" << Stride2;
VLOG(2) << "Stride1:" << Stride1;
VLOG(2) << "Stride0:" << Stride0;
VLOG(2) << "gws:[3D]" << ((new_dims[1] + 3) / 4) << " " << new_dims[3]
<< " " << (new_dims[0] * new_dims[2]);
#endif
auto& context = ctx_->As<OpenCLContext>();
CHECK(context.cl_context() != nullptr);
STL::stringstream kernel_key;
kernel_key << kernel_func_name_ << build_options_;
auto kernel = context.cl_context()->GetKernel(kernel_key.str());
int arg_idx = 0;
cl_int status = kernel.setArg(arg_idx, *x_data);
cl_int status;
status = kernel_->setArg(0, *x_data);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, *y_data);
status = kernel_->setArg(1, *y_data);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(out_H));
status = kernel_->setArg(2, out_H);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(out_W));
status = kernel_->setArg(3, out_W);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(out_C));
status = kernel_->setArg(4, out_C);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(Stride0));
status = kernel_->setArg(5, Stride0);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(Stride1));
status = kernel_->setArg(6, Stride1);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(Stride2));
status = kernel_->setArg(7, Stride2);
CL_CHECK_FATAL(status);
VLOG(2) << "gws:[3D]" << ((new_dims[1] + 3) / 4) << " " << new_dims[3]
<< " " << (new_dims[0] * new_dims[2]);
auto global_work_size =
cl::NDRange{static_cast<cl::size_type>((new_dims[1] + 3) / 4),
static_cast<cl::size_type>(new_dims[3]),
static_cast<cl::size_type>(new_dims[0] * new_dims[2])};
status = context.cl_context()->GetCommandQueue().enqueueNDRangeKernel(
kernel,
*(kernel_.get()),
cl::NullRange,
global_work_size,
cl::NullRange,
......@@ -142,6 +144,7 @@ class LayoutComputeBufferChwToImageDefault
std::string kernel_func_name_{"buffer_to_image2d"};
std::string build_options_{"-DCL_DTYPE_float"};
std::shared_ptr<cl::Event> event_{new cl::Event};
std::shared_ptr<cl::Kernel> kernel_;
};
// [ImageDefault] -> [NCHW]
......@@ -157,7 +160,7 @@ class LayoutComputeImageDefaultToBufferChw
}
VLOG(1) << "kernel_func_name_:" << kernel_func_name_;
auto& context = ctx_->As<OpenCLContext>();
context.cl_context()->AddKernel(
kernel_ = context.cl_context()->CreateKernel(
kernel_func_name_, "image/layout_kernel.cl", build_options_);
}
......@@ -191,46 +194,45 @@ class LayoutComputeImageDefaultToBufferChw
VLOG(2) << "param.y->memory_size():" << param.y->memory_size();
#endif
size_t C = new_dims[1];
size_t in_height = new_dims[2];
size_t in_width = new_dims[3];
int size_ch = in_height * in_width;
int size_block = size_ch * 4;
int size_batch = size_ch * C;
const int C = new_dims[1];
const int in_height = new_dims[2];
const int in_width = new_dims[3];
const int size_ch = in_height * in_width;
const int size_block = size_ch * 4;
const int size_batch = size_ch * C;
auto& context = ctx_->As<OpenCLContext>();
CHECK(context.cl_context() != nullptr);
STL::stringstream kernel_key;
kernel_key << kernel_func_name_ << build_options_;
auto kernel = context.cl_context()->GetKernel(kernel_key.str());
int arg_idx = 0;
cl_int status = kernel.setArg(arg_idx, *x_data);
cl_int status;
status = kernel_->setArg(0, *x_data);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(in_width));
status = kernel_->setArg(1, in_width);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(in_height));
status = kernel_->setArg(2, in_height);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, *y_data);
status = kernel_->setArg(3, *y_data);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(size_ch));
status = kernel_->setArg(4, size_ch);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(size_block));
status = kernel_->setArg(5, size_block);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(size_batch));
status = kernel_->setArg(6, size_batch);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(C));
status = kernel_->setArg(7, C);
CL_CHECK_FATAL(status);
#ifndef LITE_SHUTDOWN_LOG
VLOG(2) << "gws:[3D]" << ((new_dims[1] + 3) / 4) << " " << new_dims[3]
<< " " << (new_dims[0] * new_dims[2]);
#endif
auto global_work_size =
cl::NDRange{static_cast<cl::size_type>((new_dims[1] + 3) / 4),
static_cast<cl::size_type>(new_dims[3]),
static_cast<cl::size_type>(new_dims[0] * new_dims[2])};
status = context.cl_context()->GetCommandQueue().enqueueNDRangeKernel(
kernel,
*(kernel_.get()),
cl::NullRange,
global_work_size,
cl::NullRange,
......@@ -249,6 +251,7 @@ class LayoutComputeImageDefaultToBufferChw
std::string kernel_func_name_{"image2d_to_buffer"};
std::string build_options_{"-DCL_DTYPE_float"};
std::shared_ptr<cl::Event> event_{new cl::Event};
std::shared_ptr<cl::Kernel> kernel_;
};
// [NCHW] -> [ImageDW]
......@@ -261,7 +264,7 @@ class LayoutComputeBufferChwToImage2DNw
void PrepareForRun() override {
auto& context = ctx_->As<OpenCLContext>();
context.cl_context()->AddKernel(
kernel_ = context.cl_context()->CreateKernel(
kernel_func_name_, "buffer/layout_kernel.cl", build_options_);
}
......@@ -295,36 +298,36 @@ class LayoutComputeBufferChwToImage2DNw
auto& context = ctx_->As<OpenCLContext>();
CHECK(context.cl_context() != nullptr);
STL::stringstream kernel_key;
kernel_key << kernel_func_name_ << build_options_;
auto kernel = context.cl_context()->GetKernel(kernel_key.str());
int arg_idx = 0;
cl_int status = kernel.setArg(arg_idx, *x_data);
cl_int status;
status = kernel_->setArg(0, *x_data);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, *y_data);
status = kernel_->setArg(1, *y_data);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(out_H));
status = kernel_->setArg(2, out_H);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(out_W));
status = kernel_->setArg(3, out_W);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(out_N));
status = kernel_->setArg(4, out_N);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(Stride0));
status = kernel_->setArg(5, Stride0);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(Stride1));
status = kernel_->setArg(6, Stride1);
CL_CHECK_FATAL(status);
status = kernel.setArg(++arg_idx, static_cast<const int>(Stride2));
status = kernel_->setArg(7, Stride2);
CL_CHECK_FATAL(status);
#ifndef LITE_SHUTDOWN_LOG
VLOG(2) << "gws:[3D]" << ((out_N + 3) / 4) << " " << out_W << " "
<< (out_C * out_H);
#endif
auto global_work_size =
cl::NDRange{static_cast<cl::size_type>((out_N + 3) / 4), // N blocks
static_cast<cl::size_type>(out_W), // w
static_cast<cl::size_type>(out_C * out_H)}; // ch
status = context.cl_context()->GetCommandQueue().enqueueNDRangeKernel(
kernel,
*(kernel_.get()),
cl::NullRange,
global_work_size,
cl::NullRange,
......@@ -342,6 +345,7 @@ class LayoutComputeBufferChwToImage2DNw
std::string kernel_func_name_{"buffer_to_image2d_nw"};
std::string build_options_{"-DCL_DTYPE_float "};
std::shared_ptr<cl::Event> event_{new cl::Event};
std::shared_ptr<cl::Kernel> kernel_;
};
} // namespace opencl
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或