[WebNN EP] Cache MLTensors between runs#22278
Conversation
### Description This change enables caching `MLTensor`s between inferences runs. This is done by keeping a reference to `MLTensor`s alive after they have been released. `MLTensor`s are only destroyed once the sessions goes out of scope. ### Motivation and Context Creating and destroying `MTensor`s on every run has a non-trivial performance penalty. This performance penalty materializes when using `ort.Tensors`[location=cpu] for inputs/outputs or when using the CPU EP as a fallback EP for unsupported operators. The former could be mitigated by developer using `ort.Tensors`[location=ml-tensor]. The latter cannot be mitigated by developers.
Thanks for fixing this issue by caching MLTensors between inferences runs. Dispatching a graph with different |
|
/azp run ONNX Runtime Web CI Pipeline,Windows GPU CI Pipeline,Linux Android Emulator QNN CI Pipeline |
|
/azp run Linux CPU CI Pipeline,Linux CPU Minimal Build E2E CI Pipeline,Linux GPU CI Pipeline,Linux GPU TensorRT CI Pipeline,Linux OpenVINO CI Pipeline,Linux QNN CI Pipeline,MacOS CI Pipeline,Windows ARM64 QNN CI Pipeline,Windows CPU CI Pipeline |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
/azp run Windows GPU TensorRT CI Pipeline,onnxruntime-binary-size-checks-ci-pipeline,orttraining-linux-ci-pipeline,orttraining-linux-gpu-ci-pipeline,orttraining-ortmodule-distributed,Windows x64 QNN CI Pipeline,Big Models |
|
Azure Pipelines could not run because the pipeline triggers exclude this branch/path. |
|
/azp run Windows GPU CUDA CI Pipeline,Windows GPU DML CI Pipeline,Windows GPU Doc Gen CI Pipeline |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
Azure Pipelines could not run because the pipeline triggers exclude this branch/path. |
|
@huningxin Along with currently using |
Your plan sounds good to me. |
### Description This change enables caching `MLTensor`s between inferences runs. This is done by keeping a reference to `MLTensor`s alive after they have been released. `MLTensor`s are only destroyed once the sessions goes out of scope. ### Motivation and Context Creating and destroying `MTensor`s on every run has a non-trivial performance penalty. This performance penalty materializes when using `ort.Tensors`[location=cpu] for inputs/outputs or when using the CPU EP as a fallback EP for unsupported operators. The former could be mitigated by developer using `ort.Tensors`[location=ml-tensor]. The latter cannot be mitigated by developers.
### Description This change enables caching `MLTensor`s between inferences runs. This is done by keeping a reference to `MLTensor`s alive after they have been released. `MLTensor`s are only destroyed once the sessions goes out of scope. ### Motivation and Context Creating and destroying `MTensor`s on every run has a non-trivial performance penalty. This performance penalty materializes when using `ort.Tensors`[location=cpu] for inputs/outputs or when using the CPU EP as a fallback EP for unsupported operators. The former could be mitigated by developer using `ort.Tensors`[location=ml-tensor]. The latter cannot be mitigated by developers.
5 participants
Description
This change enables caching
MLTensors between inferences runs. This is done by keeping a reference toMLTensors alive after they have been released.MLTensors are only destroyed once the sessions goes out of scope.Motivation and Context
Creating and destroying
MTensors on every run has a non-trivial performance penalty. This performance penalty materializes when usingort.Tensors[location=cpu] for inputs/outputs or when using the CPU EP as a fallback EP for unsupported operators. The former could be mitigated by developer usingort.Tensors[location=ml-tensor]. The latter cannot be mitigated by developers.