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’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Added split op bf16/fp32 oneDNN kernel #33584

Merged
merged 11 commits into from
Jun 23, 2021
Merged

Added split op bf16/fp32 oneDNN kernel #33584

merged 11 commits into from
Jun 23, 2021

Conversation

jakpiase
Copy link
Contributor

@jakpiase jakpiase commented Jun 15, 2021 •
edited
Loading

PR types

New features

PR changes

OPs

Describe

Added split op bf16/fp32 oneDNN kernel. This PR caused 20% speed up on DPN68 model, measured on Intel(R) Core(TM) i9-9940X CPU @ 3.30GHz with 1000 repeated iterations. For bigger speed up keys optimizations are needed. For now OneDNN kernel will be used only if AxisTensor and SectionsTensorList are not set.

@paddle-bot-old
Copy link

Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

@jakpiase jakpiase marked this pull request as draft June 15, 2021 17:55
@jakpiase jakpiase marked this pull request as ready for review June 17, 2021 13:49
jczaja
jczaja previously approved these changes Jun 18, 2021
View reviewed changes
Copy link
Contributor

@jczaja jczaja left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

lidanqing-intel
lidanqing-intel previously approved these changes Jun 21, 2021
View reviewed changes
Copy link
Contributor

@lidanqing-intel lidanqing-intel left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

wozna
wozna reviewed Jun 21, 2021
View reviewed changes
Copy link
Contributor

@wozna wozna left a comment •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It would be good if you adapted these changes to the functionality of inference BF16.
This involves:

paddle/fluid/operators/split_op.cc Show resolved Hide resolved
python/paddle/fluid/tests/unittests/mkldnn/test_split_bf16_mkldnn_op.py Show resolved Hide resolved
@jakpiase jakpiase dismissed stale reviews from lidanqing-intel and jczaja via 2bfbf62 June 22, 2021 10:12
wozna
wozna approved these changes Jun 22, 2021
View reviewed changes
Copy link
Contributor

@wozna wozna left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

lidanqing-intel
lidanqing-intel approved these changes Jun 22, 2021
View reviewed changes
Copy link
Contributor

@lidanqing-intel lidanqing-intel left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.

ghost
ghost approved these changes Jun 23, 2021
View reviewed changes
Copy link

@ghost ghost left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

wanghuancoder
wanghuancoder approved these changes Jun 23, 2021
View reviewed changes
Copy link
Contributor

@wanghuancoder wanghuancoder left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@jczaja jczaja merged commit 6810650 into PaddlePaddle:develop Jun 23, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@jczaja jczaja jczaja approved these changes

@wanghuancoder wanghuancoder wanghuancoder approved these changes

@wozna wozna wozna approved these changes

@lidanqing-intel lidanqing-intel lidanqing-intel approved these changes

Assignees
No one assigned
Labels
BF16 Intel
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
5 participants
@jakpiase @jczaja @wanghuancoder @wozna @lidanqing-intel