JDK-8319979 : Regression ~13% on Renaissance-Reactors on ARM
  • Type: Bug
  • Component: core-libs
  • Sub-Component: java.util.concurrent
  • Affected Version: 22
  • Priority: P4
  • Status: Closed
  • Resolution: Duplicate
  • Submitted: 2023-11-13
  • Updated: 2023-12-15
  • Resolved: 2023-12-15
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 22
22Resolved
Related Reports
Duplicate :  
Relates :  
Description
Our ARM platform is OCI A1.160 running with numactl to node 0.

Also this change cause >3% regressions on Renaissance-Akka on Mac and Windows.

This appeared in retriage for 22-b22. I did CI build by build runs and it is due to JDK-8288899.
Comments
https://github.com/openjdk/jdk/pull/16928 should address this
06-12-2023

Also note that renaissance was updated to v15.0 last month, and among the listed changes was fixing a memory leak in reactors, that might have something to do with this. See https://github.com/renaissance-benchmarks/renaissance/releases/tag/v0.15.0
27-11-2023

One guess is that this is mainly a memory-placement/contention issue that I don't see on platforms I test on. Could you try this with the version in ongoing PR https://github.com/openjdk/jdk/pull/16725
21-11-2023

We run Renaissance 0.14.1 in the JMH form: java -jar renaissance-jmh-assembly-0.14.1.jar Reactors -t 1 -i 15 -wi 25 -f 10 For Akka: JmhAkkaUct -t 1 -i 15 -wi 25 -f 8 https://github.com/renaissance-benchmarks/renaissance/releases/download/v0.14.2/renaissance-jmh-0.14.2.jar (Can't find 0.14.1 on the web right now)
13-11-2023