Slightly improve compilation performance #2002
Merged
+69
−84
Conversation
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
This seems to allow rustc to generate better code, even though I had assumed it wouldn't make any difference.
I think I'd originally had somewhat different code in these two loops, but that has since disappeared, and there's no need -- indeed, it's inefficient -- to have two loops.
`self_copies` served two purposes: 1. Telling us when register fills have changed. 2. recording that a register is used. However, we don't make use of (2) yet, there are lots of cases where a register is used and unchanged, and `RegCopy` is a heavy-weight struct. All that means that we are putting a lot of values into `self_copies` without deriving much benefit from doing so. This commit recognises that we don't need to generate "fill changes" if fills haven't changed or if the destination fill is `Undefined`. Doing so saves about 5% of compilation of time, probably mostly because in about 90% of cases we no longer `malloc` anything.
Contributor
|
I forgot to ask if you would like to squash? |
Contributor
Author
|
No thanks. |
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Outdated suggestions cannot be applied.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.
While looking at profiles of the j2's compiler, I noticed an easy win, which this PR addresses:
rstate_diff_to_actionwas taking lots of time, mostly because itmalloced chunks of memory in cases where it didn't need to. Fixing this makes a small (~5%) improvement to compilation time.