Skip to content

Disable cuDNN 9.23.0/9.23.1 for MXFP8 attention #3173

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.

Sign up for GitHub

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

Open
cyanguwa wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Disable cuDNN 9.23.0/9.23.1 for MXFP8 attention #3173

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7a94efa
disable 9.23.0/.1 for mxfp8 attention
cyanguwa Jul 2, 2026
ad4cbef
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jul 2, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions transformer_engine/pytorch/attention/dot_product_attention/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -629,6 +629,10 @@ def _disable_all_flash_attention() -> None:
if cudnn_version < (9, 21, 0):
logger.debug("Disabling FusedAttention for MXFP8 with cuDNN < 9.21.0")
use_fused_attention = False
elif cudnn_version in ((9, 23, 0), (9, 23, 1)):
# 9.23.0/9.23.1: known bugs with MXFP8 SDPA
logger.debug("Disabling FusedAttention for MXFP8 with cuDNN 9.23.0/9.23.1")
Comment thread
cyanguwa marked this conversation as resolved.
use_fused_attention = False
elif qkv_format == "thd":
logger.debug("Disabling FusedAttention for MXFP8 with qkv_format = thd")
use_fused_attention = False
Expand Down
Loading