Specialized bytecodes#40
Merged
Hirevo merged 22 commits intoHirevo:masterfrom Feb 16, 2024
Merged
Conversation
|
Here are the benchmark results for AST interpreter
Bytecode interpreter
|
Hirevo
added
M-interpreter
|
Here are the benchmark results for AST interpreter
Bytecode interpreter
|
Contributor
Author
|
Fixed a bug which improved the perf of the send bytecodes a bit. Also added macros to avoid repetition in the bytecode handlers. These bytecode should play very nicely with #39 (especially Pretty much ready to merge in my book, just missing a couple more tests (and also for me to look at the TODOs I left in the code.. some are outdated). Let me know what you think when you get the time, Nicolas |
Hirevo
reviewed
Feb 14, 2024
Owner
There was a problem hiding this comment.
The performance increase is very nice to see for simply specializing some very common bytecode.
I really like the fact that you implemented a bytecode deduplication routine in there.
A more fleshed out "optimisation pass" system could be beneficial to cleanly implement various bytecode cleanup routines (like I thought about one for class methods that use RETURN_NON_LOCAL while outside of any closure, in which case we could transform them into simpler RETURN_LOCAL instructions).
This is something I'll add on my To-Do list.
I just have some minor remarks here and there, but nothing major.
Co-authored-by: Nicolas Polomack <nicolas@polomack.eu>
|
Here are the benchmark results for AST interpreter
Bytecode interpreter
|
|
Here are the benchmark results for
|
|
Here are the benchmark results for Bytecode interpreter
|
|
Here are the benchmark results for AST interpreter
Bytecode interpreter
|
|
Here are the benchmark results for AST interpreter
Bytecode interpreter
|
OctaveLarose
force-pushed
the
specialized-bytecodes
branch
from
24e63af to
d94bb24
Compare
|
Here are the benchmark results for
|
Contributor
Author
|
I think the test runner didn't appreciate my accidental spam. Either way, I've finished with implementing your changes and writing tests |
Hirevo
approved these changes
Feb 16, 2024
Owner
Hirevo
left a comment
Hirevo
left a comment
There was a problem hiding this comment.
I think the test runner didn't appreciate my accidental spam.
Yeah, there is no pipelining in place yet, so everything kinda ran at once.
I should have put it in place by now, sorry about that.
I'll go and implement this, it shouldn't take too much effort to build.
Either way, I've finished with implementing your changes and writing tests.
Ok, the changes look all good to me, so let's merge this !
Thanks a bunch for all this work !
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
New, specialized bytecodes:
PushConstant0PushConstant1PushConstant2Push0Push1PushNilSend1Send2Send3SendNSuperSend1SuperSend2SuperSend3SuperSendN..which were really just me looking at the PySOM bytecode set and thinking "oh yeah, that sounds like a useful bytecode"
The speedup is 9% on average or so for me.
The code duplication in
interpreter.rsfor all the specialized send bytecodes is a bit ugly, should probably replace those withsendandsuper_sendfunctions tagged with#[inline].