ARM KleidiAI Micro-Kernel: Depthwise Convolution (3x3 kernel, Stride of 1) #26654
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
damdoo01-arm
changed the title
damdoo01-arm
changed the title
Member
|
/azp run Linux QNN CI Pipeline,Win_TRT_Minimal_CUDA_Test_CI,Windows ARM64 QNN CI Pipeline,Windows GPU Doc Gen CI Pipeline |
|
Azure Pipelines successfully started running 4 pipeline(s). |
Member
|
Just FYI - I am trying to add the same for NEON in #26688 |
hariharans29
reviewed
Dec 4, 2025
| const MLAS_CONV_ALGORITHM Algorithm = Parameters->Algorithm; | ||
|
|
||
| #if defined(USE_KLEIDIAI) && !defined(_MSC_VER) | ||
| if (Algorithm == MlasConvAlgorithmExpandThenGemmSegmented && |
Member
There was a problem hiding this comment.
Maybe we can just use MlasConvAlgorithmDepthwise ? Please see #26688 ?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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
This PR introduces a dedicated kernel to perform Depthwise Convolution for specific cases (3x3 kernel, stride of 1).
Motivation and Context
We have identified opportunities for uplift for this layer type and several customer models consist of layers which would benefit from this kernel being included.