JDK-8300926 : Several startup regressions ~6-70% in 21-b6 all platforms
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 21
  • Priority: P2
  • Status: Closed
  • Resolution: Fixed
  • Submitted: 2023-01-23
  • Updated: 2024-02-12
  • Resolved: 2023-03-13
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 21
21 b14Fixed
Related Reports
Duplicate :  
JDK-8291237 - Encapsulate nmethod Deoptimization logic
Duplicate :  
JDK-8302606 - compiler/jsr292/ContinuousCallSiteTargetChange.java times out with slowdebug bits
Relates :  
JDK-8299074 - nmethod marked for deoptimization is not deoptimized
Relates :  
JDK-8302314 - JDK-8231895 may cause rough performance degradation at least for VMWare ESX Hosts/Windows Guests
Relates :  
JDK-8325613 - CTW: Massive performance regressions in recent JDKs
Relates :  
JDK-8304743 - Compile_lock and SystemDictionary updates
Relates :  
JDK-8302608 - two monitoring/stress/thread/cmon tests need a timeout adjustment
Description
Seems to be related to 
https://bugs.openjdk.org/browse/JDK-8299074

ContinuousCallSiteTargetChange.java seems to be most affected by this and may time out (see JDK-8302606 and comments in JDK-8299074).
Comments
Verified manually, by running the provided test: compiler/jsr292/ContinuousCallSiteTargetChange.java The JDK-21+13 can't run the test (times out), in JDK-21+26 the run time is acceptable. Given that our CI does run the test without failures, I consider the bug verified.
13-07-2023
A pull request was submitted for review. URL: https://git.openjdk.org/jdk17u-dev/pull/1243 Date: 2023-04-16 07:33:49 +0000
16-04-2023
Changeset: c183fce9 Author: Robbin Ehn <rehn@openjdk.org> Date: 2023-03-13 09:34:00 +0000 URL: https://git.openjdk.org/jdk/commit/c183fce9543ca15f5db632babecdb7797d0745e4
13-03-2023
Here's the logs from my jdk-21+10 stress run sightings on macosx-aarch64: $ unzip -l jdk-21+10_macosx-aarch64.8300926.zip Archive: jdk-21+10_macosx-aarch64.8300926.zip Length Date Time Name --------- ---------- ----- ---- 53466 02-17-2023 07:57 jdk-21+10_1/failures.macosx-aarch64/ContinuousCallSiteTargetChange.jtr.fastdebug 53585 02-18-2023 09:34 jdk-21+10_2/failures.macosx-aarch64/ContinuousCallSiteTargetChange.jtr.fastdebug 68717 02-19-2023 11:02 jdk-21+10_3/failures.macosx-aarch64/ContinuousCallSiteTargetChange.jtr.fastdebug --------- ------- 175768 3 files
22-02-2023
This bug is also considered to be the cause of: JDK-8302606 compiler/jsr292/ContinuousCallSiteTargetChange.java times out with slowdebug bits so the disable that I'm testing in jdk-21+10 stress testing: JDK-8302607 disable ContinuousCallSiteTargetChange.java in slowdebug mode won't need to be integrated since this bug will likely be fixed soon.
16-02-2023
A pull request was submitted for review. URL: https://git.openjdk.org/jdk/pull/12585 Date: 2023-02-16 08:38:42 +0000
16-02-2023
Discussed this with Robbin, he'll take a look. Thank you!
24-01-2023
ILW = Significant startup regressions (regression in JDK 21), with various benchmarks, no known workaround = MHH = P2
24-01-2023
Thanks for reporting, Eric. I'll take a look.
24-01-2023