You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
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.
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
PR Category
Performance Optimization
PR Types
Performance
Description
在动半
sharding optmize
时,反向计算的 grad 是partial
状态的,在做 grad clip 会分别有:grad(partial) -> allreduce(replicate) -> l2_norm
grad(partial) -> allreduce(replicate) -> multiply
在 grad clip 开始时需要融合冗余的 allreduce 通信:
replicate
的,先进行allreduce
后再进入grad clip
,通过开关FLAGS_fuse_allreduce_in_opt
控制shard
的,通过reduce_scatter
尽可能在sharding mesh
纬度切分,一般是 shard(0),其他mesh
纬度尽可能对 tensor 进行shard
,不能切分的保持replicate
,通过开关FLAGS_fuse_reducescatter_in_opt
控制在 llama7B sd8 测试吞吐(tokens/s/card)情况:
base:6153.6
开启
FLAGS_fuse_allreduce_in_opt
:6909.8(+12.2%)开启
FLAGS_fuse_reducescatter_in_opt
:7372.6(+19.7%)PCard-90131