JDK-8220204 : Windows: 3.1% regression in b8 in promo benchmark
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 13
  • Priority: P4
  • Status: Closed
  • Resolution: Cannot Reproduce
  • OS: windows
  • Submitted: 2019-03-05
  • Updated: 2019-03-27
  • Resolved: 2019-03-27
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 13
13Resolved
Related Reports
Relates :  
JDK-8217990 - C2 UseOptoBiasInlining: load of markword optimized to 0 if running with -XX:-EliminateLocks
Comments
I was unable to reproduce the regression on a Windows Server 2016 machine (105.618 ops/min both with and without the fix). Closing as Cannot Reproduce.
27-03-2019
Thanks Roland. It's very weird because I can reliably reproduce the regression (between -2.5% and -3.1%) when comparing JDK 13 b07 with and without your fix for JDK-8217990 (it shows only up with G1 though).
13-03-2019
I ran the benchmark in a windows VM with: -XX:CompileCommand=option,org.openjdk.specjvm2008.ScimarkMonteCarlo::integrate,DumpReplay then replayed compilation with and without the fix for JDK-8217990 and looked at the structure of the graph at the end of optimizations in both cases. They look identical. PrintAssembly output is also unaffected. AFAICT generated code hasn't regressed.
12-03-2019
Roland, since this was introduced by your fix for JDK-8217990, could you please have a look?
06-03-2019
Pre-ILW = Slight performance regression, with single benchmark on Windows, no workaround = MLH = P4
06-03-2019