JDK-8205722 : 6.6% regression in SPECjvm2008-MonteCarlo with G1 on Linux in JDK 11
  • Type: Bug
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: 11
  • Priority: P2
  • Status: Closed
  • Resolution: Not an Issue
  • OS: linux
  • CPU: x86
  • Submitted: 2018-06-26
  • Updated: 2018-08-24
  • Resolved: 2018-08-24
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 11
11Resolved
Related Reports
Relates :  
JDK-8191471 - Elastic TLABs for G1
Relates :  
JDK-8202662 - JFR tests fails: Could not find leak with class
Description
SPECjvm2008-MonteCarlo with G1 has regressed in JDK 11 (compared to JDK 10) by more than 6% on Linux.  There was a large regression in b12 that was mostly recovered in b13 but the score regressed again in b15.  On Windows, there was also a large regression in b12 that was recovered in b13 and the scores have since remained steady on that platform.
Comments
The regression introduced by JDK-8191471 has been fixed with JDK-8202662 in 11-b13. This has been a problem with allowing too small minimum TLABs. From my tests JDK-8202662 completely recovers the MonteCarlo performance. Still can not reproduce the b14 to b15 difference even after more reruns in our perf test infrastructure. Maybe you can help me here, [~dmcox]?
02-07-2018
The regression from b11 to b12 is caused by JDK-8191471. There were no g1 changes in b15, and no obvious gc changes that can explain the drop from b14 to b15. Needs more investigation.
29-06-2018