8315219: G1: Improve allocator pathological case where it keeps doing direct allocations instead of retiring a PLAB #15482

walulyai · 2023-08-30T08:25:15Z

Hi all,

Please review this change to improve on how G1 deals with ParallelGCBufferWastePct. Currently, any allocations larger than ParallelGCBufferWastePct of the current PLAB size are allocated directly without regard for the state of the current PLAB. This creates a pathological case where fully used up PLAB are not retired given that allocation sizes larger than ParallelGCBufferWastePct current PLAB.

In this change, we directly check whether the current PLAB can be retired i.e. the remaining PLAB space is < ParallelGCBufferWastePct of current PLAB size. This should reduce the number of direct allocations.

Testing: - mach5 Tier 1-6
- No regression was observed on perf benchmarks

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8315219: G1: Improve allocator pathological case where it keeps doing direct allocations instead of retiring a PLAB (Enhancement - P4)

Reviewers

Thomas Schatzl (@tschatzl - Reviewer) ⚠️ Review applies to f10d7f89
Albert Mingkun Yang (@albertnetymk - Reviewer)
Hamlin Li (@Hamlin-Li - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/15482/head:pull/15482
$ git checkout pull/15482

Update a local copy of the PR:
$ git checkout pull/15482
$ git pull https://git.openjdk.org/jdk.git pull/15482/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 15482

View PR using the GUI difftool:
$ git pr show -t 15482

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/15482.diff

Webrev

Link to Webrev Comment

bridgekeeper · 2023-08-30T08:26:21Z

👋 Welcome back iwalulya! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

openjdk · 2023-08-30T08:27:48Z

@walulyai The following label will be automatically applied to this pull request:

hotspot-gc

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

mlbridge · 2023-08-30T08:31:43Z

Webrevs

tschatzl

Apart from that casing issue, looks good.

tschatzl · 2023-08-30T08:59:03Z

test/hotspot/jtreg/gc/g1/plab/TestPLABPromotion.java

@@ -62,17 +63,18 @@ public class TestPLABPromotion {
    private final static String PLAB_DIRECT_ALLOCATED_FIELD_NAME = "direct allocated";
    private final static List<String> FIELDS_TO_EXTRACT = Arrays.asList(PLAB_USED_FIELD_NAME, PLAB_DIRECT_ALLOCATED_FIELD_NAME);

+    public final static int heapWordSize = Platform.is32bit() ? 4 : 8;


I would prefer to keep using the existing style for static finals.

openjdk · 2023-08-30T09:01:20Z

@walulyai This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8315219: G1: Improve allocator pathological case where it keeps doing direct allocations instead of retiring a PLAB

Reviewed-by: tschatzl, ayang, mli

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 12 new commits pushed to the master branch:

c90cd2c: 8286789: Test forceEarlyReturn002.java timed out
89d18ea: 8312018: Improve reservation of class space and CDS
dd64a4a: 8315241: (fs) Move toRealPath tests in java/nio/file/Path/Misc.java to separate JUnit 5 test
8e4cda0: 8314834: serviceability/jdwp/AllModulesCommandTest.java ignores VM flags
1ea6463: 8314835: gtest wrappers should be marked as flagless
93e82c0: 8314824: Fix serviceability/jvmti/8036666/GetObjectLockCount.java to use vm flags
7daae1f: 8314263: Signed jars triggering Logger finder recursion and StackOverflowError
6701eba: 8315117: Update Zlib Data Compression Library to Version 1.3
e29f0c2: 8297777: Convert jdk.jlink StringSharingPlugin to use Class File API
bf63945: 8298992: runtime/NMT/SummarySanityCheck.java failed with "Total committed (MMMMMM) did not match the summarized committed (NNNNNN)"
... and 2 more: https://git.openjdk.org/jdk/compare/ed1ea5fe7c6fad03ca96e7dece2127eab21a608a...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

➡️ To integrate this PR with the above commit message to the master branch, type /integrate in a new comment.

albertnetymk · 2023-08-30T10:12:02Z

src/hotspot/share/gc/g1/g1Allocator.cpp


-    PLAB* alloc_buf = alloc_buffer(dest, node_index);
-    guarantee(alloc_buf->words_remaining() <= required_in_plab, "must be");
+    guarantee(words_remaining <= required_in_plab, "must be");


What's the intend of this check? Also, if the two are equal, plab-alloc should have succeeded, so one shouldn't retire the current plab.

This is a (pre-existing) sanity check (that is correct but a little imprecise because in the == case we would not reach here) and could (and probably should because it is somewhat confusing) be removed.

Hamlin-Li

LGTM.

walulyai · 2023-08-31T08:03:23Z

Thanks for reviews @albertnetymk , @Hamlin-Li, and @tschatzl.

/integrate

openjdk · 2023-08-31T08:04:50Z

Going to push as commit 47aa6f3.
Since your change was applied there have been 17 commits pushed to the master branch:

218829e: 8315421: [BACKOUT] 8314834 serviceability/jdwp/AllModulesCommandTest.java ignores VM flags
3c8a667: 8314611: Provide more explicative error message parsing Currencies
df5e6e5: 8315248: AssertionError in Name.compareTo
3eac890: 8315061: Make LockingMode a product flag
8419a53: 8315072: Remove unneeded AdaptivePaddedAverage::operator new
c90cd2c: 8286789: Test forceEarlyReturn002.java timed out
89d18ea: 8312018: Improve reservation of class space and CDS
dd64a4a: 8315241: (fs) Move toRealPath tests in java/nio/file/Path/Misc.java to separate JUnit 5 test
8e4cda0: 8314834: serviceability/jdwp/AllModulesCommandTest.java ignores VM flags
1ea6463: 8314835: gtest wrappers should be marked as flagless
... and 7 more: https://git.openjdk.org/jdk/compare/ed1ea5fe7c6fad03ca96e7dece2127eab21a608a...master

Your commit was automatically rebased without conflicts.

openjdk · 2023-08-31T08:04:58Z

@walulyai Pushed as commit 47aa6f3.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

walulyai added 2 commits August 30, 2023 08:06

initial

734054f

rename

f10d7f8

openjdk bot added the rfr Pull request is ready for review label Aug 30, 2023

openjdk bot added the hotspot-gc hotspot-gc-dev@openjdk.org label Aug 30, 2023

tschatzl approved these changes Aug 30, 2023

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Aug 30, 2023

walulyai added 3 commits August 30, 2023 09:29

Thomas Review

e25499d

nit

f57cf96

nit

9dd2572

albertnetymk reviewed Aug 30, 2023

View reviewed changes

Albert Review

2524c68

albertnetymk approved these changes Aug 30, 2023

View reviewed changes

Hamlin-Li approved these changes Aug 30, 2023

View reviewed changes

openjdk bot added the integrated Pull request has been integrated label Aug 31, 2023

openjdk bot closed this Aug 31, 2023

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Aug 31, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8315219: G1: Improve allocator pathological case where it keeps doing direct allocations instead of retiring a PLAB #15482

8315219: G1: Improve allocator pathological case where it keeps doing direct allocations instead of retiring a PLAB #15482

walulyai commented Aug 30, 2023 •

edited by openjdk bot

Loading

bridgekeeper bot commented Aug 30, 2023

openjdk bot commented Aug 30, 2023

mlbridge bot commented Aug 30, 2023 •

edited

Loading

tschatzl left a comment

tschatzl Aug 30, 2023

openjdk bot commented Aug 30, 2023 •

edited

Loading

albertnetymk Aug 30, 2023

tschatzl Aug 30, 2023

Hamlin-Li left a comment

walulyai commented Aug 31, 2023

openjdk bot commented Aug 31, 2023

openjdk bot commented Aug 31, 2023

8315219: G1: Improve allocator pathological case where it keeps doing direct allocations instead of retiring a PLAB #15482

8315219: G1: Improve allocator pathological case where it keeps doing direct allocations instead of retiring a PLAB #15482

Conversation

walulyai commented Aug 30, 2023 • edited by openjdk bot Loading

Progress

Issue

Reviewers

Reviewing

Webrev

bridgekeeper bot commented Aug 30, 2023

openjdk bot commented Aug 30, 2023

mlbridge bot commented Aug 30, 2023 • edited Loading

Webrevs

tschatzl left a comment

Choose a reason for hiding this comment

tschatzl Aug 30, 2023

Choose a reason for hiding this comment

openjdk bot commented Aug 30, 2023 • edited Loading

albertnetymk Aug 30, 2023

Choose a reason for hiding this comment

tschatzl Aug 30, 2023

Choose a reason for hiding this comment

Hamlin-Li left a comment

Choose a reason for hiding this comment

walulyai commented Aug 31, 2023

openjdk bot commented Aug 31, 2023

openjdk bot commented Aug 31, 2023

walulyai commented Aug 30, 2023 •

edited by openjdk bot

Loading

mlbridge bot commented Aug 30, 2023 •

edited

Loading

openjdk bot commented Aug 30, 2023 •

edited

Loading