-
Notifications
You must be signed in to change notification settings - Fork 5.5k
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
Add AXPY oneDNN handler #33632
Add AXPY oneDNN handler #33632
Conversation
* Compilation of this function is possible only when Paddle is not build with CUDA nor HIP.
Thanks for your contribution! |
@jczaja @arlesniak Please review it. Now PR-CI-APPROVAL and PR-CI-OP-benchmark all passed. If you leave LGTM it will be merged. Adam's commit record keep the same I just removed skip_bf16 in last commit. |
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.
LGTM
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.
LGTM
PR types
Performance optimization
PR changes
OPs
Describe
This PR is to take over PR #33046. This PR introduces oneDNN AXPY handler providing BLAS::AXPY interface. The use of this handler is enabled at this moment only for
bfloat16
data type forsum
operator withSelectedRows
input tensors. It gives almost 10% performance improvement on CPX machine. This PR speeds up SGD operator with Selected rows with AXPY handler.