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.

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

[FlashAttention Forwad Pass] QSkipLDS with Persistent register cache policy #22

Open
wants to merge 11 commits into
base: main_old_1
Choose a base branch
from

Conversation

aska-0096
Copy link
Collaborator

@aska-0096 aska-0096 commented Oct 25, 2023

Integer initialization:
131.5T @ Batch 16 SeqQ 4096 SeqK 4096 HeadDim 128
137.2T @ Batch 64 SeqQ 4096 SeqK 4096 HeadDim 128
144.1T @ Batch 192 SeqQ 8192 SeqK 8192 HeadDim 128
Float initialization:
127.1T @ Batch 16 SeqQ 4096 SeqK 4096 HeadDim 128
132.3T @ Batch 64 SeqQ 4096 SeqK 4096 HeadDim 128
133.0T @ Batch 192 SeqQ 8192 SeqK 8192 HeadDim 128

@aska-0096 aska-0096 requested a review from asroy October 25, 2023 07:35
@aska-0096 aska-0096 self-assigned this Oct 25, 2023
@aska-0096
Copy link
Collaborator Author

Hi @asroy
Performance number updated.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant