imageUtilKernels.cu: optimize initAttentionMaskKernel#28
Open
vc-zju wants to merge 1 commit intoNVIDIA:mainfrom
Open
imageUtilKernels.cu: optimize initAttentionMaskKernel#28vc-zju wants to merge 1 commit intoNVIDIA:mainfrom
vc-zju wants to merge 1 commit intoNVIDIA:mainfrom
Conversation
…ing loop index Signed-off-by: Tiekai Bi <tiekaib@nvidia.com>
|
Would you mind sharing on which platform did you collect the performance data? |
fans-nv
approved these changes
Jan 26, 2026
|
Similar to to other community PRs, I need to cherry-pick to internal branch and then publish the change to main along with next release. |
Author
|
Sure, I collect it on jetson-Thor. I can also move this to internal gitlab if you think we need to do that. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
Type of change: Optimization
Overview:

Simply change the loop order in
imageUtilKernels.cu:initAttentionMaskKernel. Take Qwen2.5-VL as example, optimize the kernel from 2.8ms to 0.4ms for 640*640 image input.Before
After

Testing
Just use nsys to profile it and you can see it.