8387400: Force-inline Devirtualizer methods#31710
Open
shipilev wants to merge 1 commit into
Open
Conversation
|
👋 Welcome back shade! A progress list of the required criteria for merging this PR into |
|
❗ This change is not yet ready to be integrated. |
|
The total number of required reviews for this PR has been set to 2 based on the presence of this label: |
Member
Author
|
GHA failure is going to get resolved by #31707 |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The purpose for
Devirtualizeris avoid non-inlinable virtual calls to hot GC closures. It works fairly well, but at least in Shenandoah hot loops, I am seeing that theDevirtualizermethods themselves are not inlined. This poses a problem if GC-closure-calling code forces its own inlining, the hot GC closure targets force their own inlining, butDevirtualizerthat sits in the middle is not. In this case, compilers seem to use this "opportunity" to cut the aggressive inlining off in the middle. It IMO makes sense to force-inlineDevirtualizermethods themselves as well.This improves Shenandoah marking times for about +2..10% in hot benchmarks. I measured only Shenandoah, because this was a part of larger project and I had marking loop tests ready. But I suppose it will benefit things generically. I can see no ill effects on build times, the difference is a wash in the noise.
The downside is a bump in binary size. With current version, fastdebug
libjvm.sosize is +0.6% larger, releaselibjvm.sois +0.4% larger. I presume this is a fair price to pay for a palpably faster GC code.Additional testing:
libjvm.sosizesProgress
Issue
Reviewing
Using
gitCheckout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/31710/head:pull/31710$ git checkout pull/31710Update a local copy of the PR:
$ git checkout pull/31710$ git pull https://git.openjdk.org/jdk.git pull/31710/headUsing Skara CLI tools
Checkout this PR locally:
$ git pr checkout 31710View PR using the GUI difftool:
$ git pr show -t 31710Using diff file
Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/31710.diff
Using Webrev
Link to Webrev Comment