perf: accelerate JIT compilation speed #618
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
|
Using nanobind or ctypes/cython for ffi could make the compilation speed even faster, but requires some fundamental changes to the codebase, I prefer sticking to pybind at this moment. |
yzh119
added a commit
that referenced
this pull request
Nov 20, 2024
Some of the commits for fixing MLA are missing in #618, this PR add them back.
yzh119
added a commit
that referenced
this pull request
Nov 20, 2024
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Outdated suggestions cannot be applied.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.
Current JIT compilation is slow because we rely on a huge header
<torch/extension.h>which is too heavy for our use case.This PR refactors the codebase to only include necessary headers for pybind, and moves most of torch runtime API calls from C++ to python.
The compilation time was reduced from 48 seconds to 18 seconds for lightweight operators such as norm.