saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-04 04:07:22 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

TRT EP memory leak fix (#7415)

Browse source 
* fix memory leak

* small refactor

* code refactor
This commit is contained in:
Chi Lo 2021-04-23 12:04:23 -07:00 committed by GitHub
parent 043883b52d
commit f1c3f3fcc1
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 4 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
onnxruntime/core/providers/tensorrt/tensorrt_execution_provider.cc
View file

@ -473,7 +473,7 @@ TensorrtExecutionProvider::TensorrtExecutionProvider(const TensorrtExecutionProv
throw std::runtime_error("Failed to create directory " + cache_path_);
}
}
runtime_ = nvinfer1::createInferRuntime(GetTensorrtLogger());
runtime_ = tensorrt_ptr::unique_pointer<nvinfer1::IRuntime>(nvinfer1::createInferRuntime(GetTensorrtLogger()));
}
const std::string engine_decryption_enable_env = onnxruntime::GetEnvironmentVar(tensorrt_env_vars::kDecryptionEnable);
@ -1243,7 +1243,7 @@ common::Status TensorrtExecutionProvider::Compile(const std::vector<Node*>& fuse
&engines_[context->node_name], &contexts_[context->node_name], &builders_[context->node_name],
&networks_[context->node_name], input_info_[context->node_name], output_info_[context->node_name],
input_shape_ranges_[context->node_name], &tensorrt_mu_, &fp16_enable_, &int8_enable_, &max_workspace_size_,
trt_node_name_with_precision, engine_cache_enable_, cache_path_, runtime_, nullptr,
trt_node_name_with_precision, engine_cache_enable_, cache_path_, runtime_.get(), nullptr,
allocator_, dynamic_range_map, engine_decryption_enable_, engine_decryption_};
*state = p.release();
return 0;
@ -1295,9 +1295,8 @@ common::Status TensorrtExecutionProvider::Compile(const std::vector<Node*>& fuse
engine_file.seekg(0, std::ios::beg);
std::unique_ptr<char[]> engine_buf{new char[engine_size]};
engine_file.read((char*)engine_buf.get(), engine_size);
auto runtime_ = trt_state->runtime;
*(trt_state->engine) = tensorrt_ptr::unique_pointer<nvinfer1::ICudaEngine>(
runtime_->deserializeCudaEngine(engine_buf.get(), engine_size, nullptr));
trt_state->runtime->deserializeCudaEngine(engine_buf.get(), engine_size, nullptr));
if (trt_state->engine == nullptr) {
return ORT_MAKE_STATUS(ONNXRUNTIME, EP_FAIL, "TensorRT EP Failed to Build Engine.");
}

2
onnxruntime/core/providers/tensorrt/tensorrt_execution_provider.h
View file

@ -149,7 +149,7 @@ class TensorrtExecutionProvider : public IExecutionProvider {
bool dump_subgraphs_ = false;
bool engine_cache_enable_ = false;
std::string cache_path_;
nvinfer1::IRuntime* runtime_ = nullptr;
tensorrt_ptr::unique_pointer<nvinfer1::IRuntime> runtime_ = nullptr;
OrtMutex tensorrt_mu_;
int device_id_;
AllocatorPtr allocator_;