-
Notifications
You must be signed in to change notification settings - Fork 75
[Merge] Reland d36e579 from upstream. #2785
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
|
Please fix the UT failures. |
| if (auto sliceLayout = mlir::dyn_cast<SliceEncodingAttr>(layout)) { | ||
| threadsPerWarp = getThreadsPerWarp(sliceLayout.getParent()); | ||
| threadsPerWarp.erase(threadsPerWarp.begin() + sliceLayout.getDim()); | ||
| warpsPerCTA = getWarpsPerCTA(sliceLayout.getParent()); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is a real bug of the upstream. But the NV and AMD code doesn't uses this legacy layout function.
Need to align our lowering code to the upstream once we enabled the LL for DPAS related layout.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Please create an issue to track this.
5ea8ec8 to
9a0e517
Compare
whitneywhtsang
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Please do not squash and merge.
…erface in DPAS layout.
9a0e517 to
1aebd43
Compare
Reland d36e579 that is temporarily reverted in #2783 by 016ee00.
The PR based on PR #2766