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

Remove vectorized for those data type with size greater or euqal to double and reorganize the codes to decrease the size of the static library. #57064

Merged
merged 11 commits into from
Sep 11, 2023

Conversation

Xreki
Copy link
Contributor

@Xreki Xreki commented Sep 7, 2023

PR types

Others

PR changes

OPs

Description

Pcard-70459

ElementwiseKernelBroadcastKernel中使用了向量化技术,优化显存访问吞吐。如采用vec_size=4,float性能可提升约5%,float16性能可提升约30%,但是对于double等64位以上的数据类型,则没有明显的性能提升。故本PR避免对64位以上数据类型做向量化,减少编译静态库的体积。单个Ampere架构

API-Benchmark Fail应与本PR无关,在https://xly.bce.baidu.com/paddlepaddle/paddle/newipipe/detail/9126824/job/23820104 也观察到呈现同样的现象。

@paddle-bot
Copy link

paddle-bot bot commented Sep 7, 2023

你的PR提交成功,感谢你对开源项目的贡献!
请关注后续CI自动化测试结果,详情请参考Paddle-CI手册
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

@Xreki Xreki changed the title Reorganize the codes to decrease the size of the static library. Remove vectorized for those data type with size greater or euqal to double and reorganize the codes to decrease the size of the static library. Sep 11, 2023
@Xreki Xreki merged commit 723f22d into PaddlePaddle:develop Sep 11, 2023
@Xreki Xreki deleted the op/opt_codes branch September 11, 2023 07:22
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.

None yet

3 participants