JDK-7143341 : +TieredCompilation regression on 7u4-b10 on Solaris
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.
Observing a performance regression with 7u4-b10 with +TieredCompilation
Code cache sweeper must yield to safepoint. With high numbers of methods sweeping can block the safepoint from happening. See CR 7145345.
The regression seems to have many parts:
Most of the regression was caused by Parallel Scavenger not cleaning up the scavenge roots in code list after young GCs. In addition it's more profitable to promote the objects directly pointed by the code immediately instead of ping-ponging them in the survivor spaces, this will also allow to cleanup the roots in code list as soon as possible. This will be fixed in 7144296: PS: Optimize nmethods processing
Disable tiered comilation, -XX:-TieredCompilation