fix singlepass slow compilation of many function wasm by using dynasm VecAssembler #2681
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
syrusakbary
approved these changes
Nov 15, 2021
|
bors r+ |
bors bot
added a commit
that referenced
this pull request
Nov 15, 2021
2681: fix singlepass slow compilation of many function wasm by using dynasm VecAssembler r=syrusakbary a=ailisp <!-- Prior to submitting a PR, review the CONTRIBUTING.md document for recommendations on how to test: https://github.com/wasmerio/wasmer/blob/master/CONTRIBUTING.md#pull-requests --> # Description <!-- Provide details regarding the change including motivation, links to related issues, and the context of the PR. --> We found in certain benchmark (a wasm with 150_000 no-op functions), compilation is slowed down by dynasmrt's mmapand mprotect syscalls, however, those mmap isn't used in wasmer and can be replaced by simpler VecAssembler. In this [bench](near/wasmer#71) it's a 10x speedup in singlepass backend compilation time. # Review - [ ] Add a short description of the change to the CHANGELOG.md file Co-authored-by: Bo Yao <[email protected]>
|
Build failed: |
|
@ailisp can you do a |
|
bors r+ |
nagisa
referenced
this pull request
in near/wasmer
Nov 24, 2021
2681: fix singlepass slow compilation of many function wasm by using dynasm VecAssembler r=ptitSeb a=ailisp <!-- Prior to submitting a PR, review the CONTRIBUTING.md document for recommendations on how to test: https://github.com/wasmerio/wasmer/blob/master/CONTRIBUTING.md#pull-requests --> <!-- Provide details regarding the change including motivation, links to related issues, and the context of the PR. --> We found in certain benchmark (a wasm with 150_000 no-op functions), compilation is slowed down by dynasmrt's mmapand mprotect syscalls, however, those mmap isn't used in wasmer and can be replaced by simpler VecAssembler. In this [bench](#71) it's a 10x speedup in singlepass backend compilation time. - [ ] Add a short description of the change to the CHANGELOG.md file Co-authored-by: Bo Yao <[email protected]>
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.
Description
We found in certain benchmark (a wasm with 150_000 no-op functions), compilation is slowed down by dynasmrt's mmapand mprotect syscalls, however, those mmap isn't used in wasmer and can be replaced by simpler VecAssembler. In this bench it's a 10x speedup in singlepass backend compilation time.
Review