This repository has been archived by the owner on Aug 30, 2024. It is now read-only.
Optimization of Layernormalization #103
Merged
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
luoyu-intel
requested review from
airMeng and
zhewang1-intc
and removed request for
airMeng
airMeng
approved these changes
Jan 31, 2024
There was a problem hiding this comment.
I believe @zhewang1-intc implemented gemm+layernorm fusion before, is it necessary?
| @@ -4,7 +4,7 @@ project(bestla LANGUAGES CXX VERSION 0.1.0) | |||
| file(GLOB headers ${PROJECT_NAME}/*.h ${PROJECT_NAME}/*.hpp) | |||
| file(GLOB xbyak_headers ${PROJECT_NAME}/xbyak/*.h ${PROJECT_NAME}/xbyak/*.hpp) | |||
|
|
|||
| option(BTLA_USE_OPENMP "Enable OpenMP thread pool" ON) | |||
| option(BTLA_USE_OPENMP "Enable OpenMP thread pool" OFF) | |||
There was a problem hiding this comment.
we already have a customized threadpool implemented?
There was a problem hiding this comment.
It's better not to set it ON as default. It can be set in neural_speed as it uses OMP as default.
Yes, it's for ONNX's definition. It can have scale and bias. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Type of Change
Implement AVX2 and AVX512F Layernormalization in BesTLA.
Use BesTLA kernel if src0 is contiguous.
~3x speedup.
Before:
GPTJ: perf_total_per_op_us[ NORM] = 0.477 ms
LLAMA2: perf_total_per_op_us[ RMS_NORM] = 0.362 ms
After:
GPTJ: perf_total_per_op_us[ NORM] = 0.055 ms
LLAMA2: perf_total_per_op_us[ RMS_NORM] = 0.117 ms