JDK-8295069 : [PPC64] Performance regression after JDK-8290025
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 20
  • Priority: P3
  • Status: Resolved
  • Resolution: Fixed
  • CPU: ppc
  • Submitted: 2022-10-10
  • Updated: 2022-10-17
  • Resolved: 2022-10-11
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 20
20 b19Fixed
Related Reports
Relates :  
Description
JDK-8290025 has enabled usage of nmethod entry barriers regardless of the selected GC. This causes a noticeable performance drop (about 3% in SPEC jvm98, jvm2008, and a bit more in jbb2015). PPC64 suffers more than x86 because the nmethod entry barrier contains a memory barrier for GCs which patch nmethods during a concurrent phase.

However, not all GCs perform patching of nmethods while the Java Threads are running, so, the memory barrier is not always needed.
Comments
Changeset: 945950d8 Author: Martin Doerr <mdoerr@openjdk.org> Date: 2022-10-11 19:19:41 +0000 URL: https://git.openjdk.org/jdk/commit/945950d863ebe984e099d83f967adce71892bb95
11-10-2022

ILW = performance regression; 3 benchmarks; no workaround = MMH = P3
10-10-2022

A pull request was submitted for review. URL: https://git.openjdk.org/jdk/pull/10632 Date: 2022-10-10 14:02:09 +0000
10-10-2022