[webgpu] add usage of ReadonlyAllocator for WebGPU EP #25690
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
There was a problem hiding this comment.
Pull Request Overview
This PR integrates the WebGPU execution provider with the ReadonlyAllocator pattern introduced in PR #25348, allowing separate allocators for initializers versus regular operations. The change makes the allocator type depend on OrtAllocatorType rather than session initialization status.
Key changes:
- Creates two separate allocators in
CreatePreferredAllocators()- one for read-only initializers and one for regular operations - Adds a dedicated initializer buffer manager to
WebGpuContext - Simplifies the
BufferManager::Create()method by consolidating UMA buffer creation logic
Reviewed Changes
Copilot reviewed 8 out of 8 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| webgpu_execution_provider.h | Removes session initialization callback and allocator member variable |
| webgpu_execution_provider.cc | Updates allocator creation to use two separate allocators and removes session initialization handling |
| webgpu_context.h | Adds initializer buffer manager accessor and documentation |
| webgpu_context.cc | Creates dedicated initializer buffer manager with disabled caching |
| buffer_manager.h | Removes CreateUMA method and simplifies Create method signature |
| buffer_manager.cc | Consolidates UMA buffer creation logic into the main Create method |
| allocator.h | Updates constructor to accept read-only flag and removes session initialization method |
| allocator.cc | Implements read-only allocator logic and consolidates buffer creation |
guschmue
approved these changes
Aug 11, 2025
gedoensmax
pushed a commit
to gedoensmax/onnxruntime
that referenced
this pull request
Sep 2, 2025
### Description add usage of `ReadonlyAllocator` for WebGPU EP ### Motivation and Context `ReadonlyAllocator` is added in microsoft#25348 to allow an EP to register a separated allocator only use for initializers. WebGPU EP already has different handling to initializers and non-initializers, and this change makes WebGPU EP to use the preferred way to deal with it. Now the allocator depends on the `OrtAllocatorType` instead of the session initialization status.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
add usage of
ReadonlyAllocatorfor WebGPU EPMotivation and Context
ReadonlyAllocatoris added in #25348 to allow an EP to register a separated allocator only use for initializers.WebGPU EP already has different handling to initializers and non-initializers, and this change makes WebGPU EP to use the preferred way to deal with it. Now the allocator depends on the
OrtAllocatorTypeinstead of the session initialization status.