Feature/graph net bench

[leon062112]

PR Category

Feature Enhancement

Description

Add SampleRemoteExecutor

[paddle-bot]

Thanks for your contribution!

[Copilot]

Pull request overview

This PR adds remote execution capabilities for graph network benchmarking by implementing a gRPC-based remote executor. The feature allows test execution on a remote server via the SampleRemoteExecutor class, with corresponding client and server implementations.

Changes:

• Implements SampleRemoteExecutor with gRPC client functionality for compressing/sending model directories and receiving results
• Adds test_remote_reference_device.py client to execute reference device tests remotely
• Updates gRPC server to handle model execution requests with tarball compression/decompression
• Integrates remote testing into the subgraph decomposition workflow with new test module configuration

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated 23 comments.

Show a summary per file

File	Description
graph_net/torch/test_remote_reference_device.py	New client for executing reference device tests on remote servers via gRPC
graph_net/graph_net_bench/sample_remote_executor.py	Complete implementation of remote executor with compression and gRPC communication
graph_net/graph_net_bench/grpc/server.py	Server implementation for handling remote execution requests with subprocess execution
graph_net/graph_net_bench/grpc/sample_remote_executor_test.py	Test/CLI client for the remote executor (has critical bugs)
graph_net/test/subgraph_decompose_and_evaluation_step_test.sh	Adds configuration for test_remote_reference_device module
graph_net/subgraph_decompose_and_evaluation_step.py	Integrates remote reference device testing into task scheduler
graph_net/graph_net_bench/grpc/message.proto	Adds random_seed field to ExecutionRequest
graph_net/graph_net_bench/grpc/message_pb2.py	Regenerated protobuf Python code with updated schema
graph_net/graph_net_bench/grpc/message_pb2_grpc.py	Regenerated gRPC Python code with formatting changes
graph_net/graph_net_bench/grpc/client.py	Removed old example client code
graph_net/graph_net_bench/grpc/init.py	Added package marker with comment

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The tar.extractfile(member).read() doesn't explicitly close the file object returned by extractfile(). While it may be garbage collected, it's better practice to use a context manager or explicitly close the file to avoid resource leaks, especially in a server that may handle many requests.

Suggested change

	file_content = tar.extractfile(member).read()
	extracted_file = tar.extractfile(member)
	if extracted_file is None:
	continue
	with extracted_file as f:
	file_content = f.read()

[Copilot]

The method signature indicates it returns 'dict' but lacks documentation about the structure of the dictionary. Add a docstring explaining that it returns a dict mapping filenames (str) to file contents (bytes) to help users understand the return value structure.

Suggested change

	def __call__(self, model_path: str, random_seed: int) -> dict:
	def __call__(self, model_path: str, random_seed: int) -> dict:
	"""
	Execute the remote model with the given seed and return its output files.
	Returns:
	dict: A mapping from output filenames (str) to their file contents (bytes).
	"""