Wint8 gemm and gemv opt #59291
Merged
Wint8 gemm and gemv opt #59291
Conversation
This file contains 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! |
wwbitejotunn
force-pushed
the
develop_wint8_gemm_opt
branch
from
74589d4 to
1402d7c
Compare
|
私以为，升级点在PR里可能应该更详细一点～ |
zhoutianzi666
previously approved these changes
Dec 1, 2023
| @@ -5163,7 +5163,7 @@ void WeightQuantizeInferMeta(const MetaTensor& x, | |||
| out->set_dtype(DataType::INT8); | |||
|
|
|||
| scale->set_dims(phi::make_ddim(dim_scale)); | |||
| scale->set_dtype(DataType::FLOAT32); | |||
There was a problem hiding this comment.
对的, 这边scale改为了bf16/fp16, 能够有更好的性能, 精度应该也能保证, 已经同步修改了weight_quant op中计算scale的部分, 以及paddlenlp代码中scale权重初始化的逻辑
已添加了升级点和测试数据~ |
heavengate
previously approved these changes
Dec 4, 2023
raindrops2sea
previously approved these changes
Dec 4, 2023
wwbitejotunn
dismissed stale reviews from Xreki, raindrops2sea, heavengate, and zhoutianzi666
via
496a120
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 types
Performance optimization
PR changes
OPs
Description
Pcard-71501
This PR optimize the speed of weight only gemm and gemv gpu kernel.
To speed up the weight-only gemm, following features were adopted
For gemms with problem sizes in llama13b, we obtain a 1.34x gemm kernel speed in A100 80G.