Skip to content
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鈥檒l occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

2.2.0+ regresses SDPA performance on Windows #125070

Open
Xemorr opened this issue Apr 26, 2024 · 0 comments
Open

2.2.0+ regresses SDPA performance on Windows #125070

Xemorr opened this issue Apr 26, 2024 · 0 comments
Labels
module: multi-headed-attention module: windows Windows support for PyTorch triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@Xemorr
Copy link

Xemorr commented Apr 26, 2024 •
edited

馃悰 Describe the bug

In PyTorch <2.2.0, SDPA (scaled_dot_product_attention) supports Flash Attention v1 on Windows. In PyTorch 2.2.0>=, it does not support any Flash Attention on Windows.

Versions

This is a report of a regression between 2.1.2 and 2.2.0+

cc @peterjc123 @mszhanyi @skyline75489 @nbcsm @vladimir-aubrecht @iremyux @Blackhex @cristianPanaite
@mikaylagawarecki mikaylagawarecki added module: windows Windows support for PyTorch module: multi-headed-attention triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module labels Apr 26, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
module: multi-headed-attention module: windows Windows support for PyTorch triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@Xemorr @mikaylagawarecki