8266073: Regression ~2% in Derby after 8261804

[lkorinth]

My change in 8261804 made processing_is_mt() to always be false when _active_mt_degree is one. This is better because we then need not to execute on another thread. However, at a few places we do not change the _active_mt_degree if processing_is_mt() is false. We could then be stuck at 1 thread. I now change the code to always change the mt degree.

Derby does no longer regress for g1, and tier 1 passes. I am running tier 1-3 at the moment.

---

Progress

• Change must not contain extraneous whitespace
• Commit message must refer to an issue
• Change must be properly reviewed

Issue

• JDK-8266073: Regression ~2% in Derby after 8261804

Reviewers

• Albert Mingkun Yang (@albertnetymk - Committer) ⚠️ Review applies to c8bf51f
• Stefan Johansson (@kstefanj - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.java.net/jdk pull/3972/head:pull/3972
$ git checkout pull/3972

Update a local copy of the PR:
$ git checkout pull/3972
$ git pull https://git.openjdk.java.net/jdk pull/3972/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 3972

View PR using the GUI difftool:
$ git pr show -t 3972

Using diff file

Download this PR as a diff file:
https://git.openjdk.java.net/jdk/pull/3972.diff

[bridgekeeper]

👋 Welcome back lkorinth! 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]

@lkorinth 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]

Webrevs

• 02: Full - Incremental (2bc42aee)
• 01: Full - Incremental (35b602d9)
• 00: Full (c8bf51f9)

[kstefanj]

Looks good. Just one question, did you do any analysis of the G1 reference processing during concurrent mark? In that code path the call to set_active_mt_degree() comes after checking processing_is_mt(). The comment suggests this is ok, just wanted to make sure this is still true after JDK-8261804:

jdk/src/hotspot/share/gc/g1/g1ConcurrentMark.cpp

Line 1585 in 995e956

rp->set_active_mt_degree(active_workers);

[lkorinth]

Good catch Stefan, I will fix it.

[lkorinth]

My fix is using ParallelRefProcEnabled instead of rp->processing_is_mt() as rp->processing_is_mt() can return false even if ParallelRefProcEnabled is true if only one thread is set at the moment. I also set the thread-count directly after calculating it and then use rp->processing_is_mt() to decide if we should use a work gang (we should not use a work gang if only one thread is used even if ParallelRefProcEnabled is true.

[kstefanj]

Looks good.

[openjdk]

@lkorinth 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:

8266073: Regression ~2% in Derby after 8261804

Reviewed-by: ayang, sjohanss

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 35 new commits pushed to the master branch:

• 2568d18: 8267047: Put serviceability/sa/TestJmapCoreMetaspace.java back on ZGC problem list due to JDK-8267045
• accbfea: 8226216: parameter modifiers are not visible to javac plugins across compilation boundaries
• 69daedf: 8266845: Shenandoah: Simplify SBS::load_reference_barrier implementation
• 7433821: 8250658: Performance of ClipFlatOval Renderperf test is very low
• 4727187: 8266567: Fix javadoc tag references in sun.management.jmxremote.ConnectorBootstrap
• 11759bf: 8266798: C1: More types of instruction can also apply LoopInvariantCodeMotion
• dcf250d: 8233378: CHT: Fast reset
• f3b510b: 8266923: [JVMCI] expose StackOverflow::_stack_overflow_limit to JVMCI
• 548899d: 8266189: Remove C1 "IfInstanceOf" instruction
• b46086d: 8266874: Clean up C1 canonicalizer for TableSwitch/LookupSwitch
• ... and 25 more: https://git.openjdk.java.net/jdk/compare/9e6e2228cba05ff2ee3a4014a0a92bdd08d016d9...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.

[lkorinth]

Thanks Albert and Stefan!

[lkorinth]

/integrate

[openjdk]

@lkorinth Since your change was applied there have been 72 commits pushed to the master branch:

• 02f895c: 8252685: APIs that require JavaThread should take JavaThread arguments
• 2066f49: 8266764: [REDO] JDK-8255493 Support for pre-generated java.lang.invoke classes in CDS dynamic archive
• 8c71144: 8265153: add time based test for ThreadMXBean.getThreadInfo() and ThreadInfo.getLockOwnerName()
• 10cafd2: 8267153: Problemlist jdk/jfr/event/gc/collection/TestG1ParallelPhases.java to remove the noise from CI
• f3fb5a4: 8266942: gtest/GTestWrapper.java os.iso8601_time_vm failed
• 7ab6dc8: 6676643: Improve current C_GetAttributeValue native implementation
• 28f1c7a: 8249395: (macos) jpackage tests timeout on MacPro5_1 systems
• be0a655: 8254598: StringDedupTable should use OopStorage
• 360928d: 8260046: Assert left >= right in pointer_delta() methods
• 5eda812: 8267180: Typo in copyright header for HashesTest
• ... and 62 more: https://git.openjdk.java.net/jdk/compare/9e6e2228cba05ff2ee3a4014a0a92bdd08d016d9...master

Your commit was automatically rebased without conflicts.

Pushed as commit f422787.

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