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

[cherry-pick] temporarily fix the performance drop of recurrent op #36053

Merged
merged 1 commit into from
Sep 25, 2021
Merged

[cherry-pick] temporarily fix the performance drop of recurrent op #36053

merged 1 commit into from
Sep 25, 2021

Conversation

baoachun
Copy link
Contributor

PR types

Bug fixes

PR changes

OPs

Describe

修改recurrent算子内部线程锁的位置,以提高该算子性能。

@paddle-bot-old
Copy link

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

Copy link
Contributor

@jiweibo jiweibo left a comment

Choose a reason for hiding this comment

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

LGTM

贴下性能数据对比吧

@baoachun
Copy link
Contributor Author

CreateStepScopes前面加锁
图片

CreateStepScopes内部加锁
图片

去掉锁
图片

可以看到将锁移到CreateStepScopes内部能大幅提升网络性能,不过还需要确认能否直接去掉锁。

@Superjomn Superjomn merged commit 33fbdaf into PaddlePaddle:release/2.2 Sep 25, 2021
@LeoMax-Xiong
Copy link

CreateStepScopes前面加锁 图片

CreateStepScopes内部加锁 图片

去掉锁 图片

可以看到将锁移到CreateStepScopes内部能大幅提升网络性能,不过还需要确认能否直接去掉锁。

这个是否是在多线程情况下,比单线程的效率有一倍的下降??

@baoachun
Copy link
Contributor Author

CreateStepScopes前面加锁 图片
CreateStepScopes内部加锁 图片
去掉锁 图片
可以看到将锁移到CreateStepScopes内部能大幅提升网络性能,不过还需要确认能否直接去掉锁。

这个是否是在多线程情况下,比单线程的效率有一倍的下降??

多线程效率下降是正常的,不一定都是因为这个锁的原因

@baoachun baoachun deleted the fix_recurrent_performance branch December 31, 2021 07:27
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.

4 participants