Cherry-pick release:1.23.0 PRs to rel-1.23.0 #25985
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
The "custom nuget pipeline" and QNN Nuget pipeline do not support adding "rc" postfix to the nuget package's version string. This PR fixes that. (cherry picked from commit 234cc63)
Currently our macOS wheels use Metadata-Version 2.1 Currently our Windows and Linux wheels use Metadata-Version 2.4 Because they use different versions of the ["wheel"](https://pypi.org/project/wheel/#history) package that is used for generating the wheel files. When publishing the macOS wheels to Azure DevOps feed, I got the following error: ``` ERROR: Failed to upload onnxruntime-1.23.0.dev20250903002-cp310-cp310-macosx_13_0_universal2.whl. Return code: 1 STDOUT: Uploading distributions to https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ORT-Nightly/pypi /upload ERROR InvalidDistribution: Invalid distribution metadata: dynamic introduced in metadata version 2.2, not 2.1 ``` This PR makes them consistent. (cherry picked from commit 4e2699f)
### Description <!-- Describe your changes. --> OrtMemoryInfo has a name which is pointer. With the recent changes a user can pass an arbitrary name and then deallocate the string as the API does not require it. Make the name a `std::strng` to own it and refactor. ### Motivation and Context <!-- - Why is this change required? What problem does it solve? - If it fixes an open issue, please link to the issue here. --> We will have a dangling pointer especially when it comes to other languages. The way it used to work in the past, is that the string would be compared to an internal constant and that constant would be used then. (cherry picked from commit 8ad0614)
### Description Fixes memory leak in `OrtEpFactory::GetSupportedDevices()`. The `OrtKeyValuePairs` instance created by the factory was not released. This memory leak can be reproduced by running the unit test [QnnHTPBackendTests.AutoEp_PreferNpu](https://github.com/microsoft/onnxruntime/blob/4e2699fbf24d96d2a0261b30509864da08b701de/onnxruntime/test/providers/qnn/qnn_basic_test.cc#L1426). ### Motivation and Context Fix memory leak that one encounters when using automatic EP selection (e.g., PREFER_NPU) with QNN EP. (cherry picked from commit 8570298)
…ts buffer (#25971) ### Description The memory alignment for the pre-packed weights buffer was accidentally changed for 8-bit Gemms on x86 while supporting the ARM64 equivalent 8-bit Gemm kernel in #25110. This change in alignment could either cause perf penalty or seg-fault depending on the platform while the corresponding aligned data load instruction is executed in the Gemm kernel. This changes fixes it as well as adds back a couple of tests to the MLAS 8-bit Gemm test suite and fixes a minor nit in the test file. ### Motivation and Context Resolve packaging pipeline crash (cherry picked from commit 96f4595)
adrianlizarraga
approved these changes
Sep 8, 2025
This was referenced Sep 8, 2025
Closed
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.
This PR cherry-picks the following PRs to the rel-1.23.0 branch: