Repository navigation
Fix GatherBlockQuantized default zero point on CUDA - #31693
Merged
Tianlei Wu (tianleiwu) merged 3 commits intoAug 10, 2026
Merged
Conversation
|
Azure Pipelines: There may be pipelines that require an authorized user to comment /azp run to run. |
Co-authored-by: justinchuby <11205048+justinchuby@users.noreply.github.com>
Copilot
AI
changed the title
[WIP] Fix GatherBlockQuantized default zero point discrepancy between CPU and CUDA
Fix GatherBlockQuantized default zero point on CUDA
Aug 6, 2026
Tianlei Wu (tianleiwu)
previously approved these changes
Aug 6, 2026
Justin Chu (justinchuby)
marked this pull request as ready for review
August 6, 2026 18:22
Justin Chu (justinchuby)
enabled auto-merge (squash)
August 6, 2026 18:22
auto-merge was automatically disabled
August 7, 2026 02:26
Pull request was closed
|
Azure Pipelines: There may be pipelines that require an authorized user to comment /azp run to run. |
Tianlei Wu (tianleiwu)
enabled auto-merge (squash)
August 7, 2026 02:26
Contributor
|
Copilot, please fix CI pipeline failures. Some failed test cases need update expected results. |
Co-authored-by: tianleiwu <30328909+tianleiwu@users.noreply.github.com>
Contributor
Author
Tianlei Wu (tianleiwu)
approved these changes
Aug 10, 2026
Ti-Tai Wang (titaiwangms)
approved these changes
Aug 10, 2026
Tianlei Wu (tianleiwu)
deleted the
copilot/fix-default-zero-point-issue
branch
August 10, 2026 16:11
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
uint8zero points to2^(bits - 1)in the CUDA kernel.(10 - 8) * scalefor0xAAweights.Motivation and Context
CUDA previously treated omitted
zero_pointsas zero, while CPU used the documented midpoint default. This caused execution-provider-dependent output and corrupted quantized embedding rows.