Fp8 block scale quantization for fmha fwd #3330
Open
+667
−84
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
ltqin
requested review from
ThomasNing,
afagaj,
andriy-ca,
aosewski,
asleepzzz,
bartekxk,
carlushuang,
cgmillette,
coderfeli,
geyyer,
illsilin,
poyenc,
qianfengz,
shumway,
tenpercent and
vidyasagar-amd
as code owners
This reverts commit 356c3c9.
poyenc
reviewed
Jan 15, 2026
poyenc
reviewed
Jan 15, 2026
poyenc
reviewed
Jan 16, 2026
poyenc
reviewed
Jan 16, 2026
poyenc
reviewed
Jan 16, 2026
poyenc
reviewed
Jan 16, 2026
poyenc
reviewed
Jan 16, 2026
poyenc
reviewed
Jan 16, 2026
Contributor
|
LGTM. Please make this PR won't introduce performance regression before mergin it. |
poyenc
previously approved these changes
Jan 19, 2026
Collaborator
|
Please let's make sure we don't break AITER with these changes! |
Contributor
|
@ltqin could you update the CHANGELOG.md as well? |
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Outdated suggestions cannot be applied.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.
Proposed changes
Implement fp8 block scale quantization for fmha fwd
Checklist
./build/bin/tile_example_fmha_fwd -init=3 -b=16 -s=256 -s_k=1024 -h=16 -h_k=1 -d=128 -prec=fp8bf16 -vlayout=r -qscale=2 -kname=1 -mode=0 -v=1 -operm=1 -iperm=1