-
Enhancement
-
Resolution: Duplicate
-
P4
-
23
In `PSParallelCompact::mark_obj`, `_summary_data.add_obj` uses atomic-operation to record live bytes info, which hinders scalability during Full GC.
- duplicates
-
JDK-8325553 Parallel: Use per-marker cache for marking stats during Full GC
- Resolved
- relates to
-
JDK-8320165 Parallel: Full GC code is very slow due to quadratic calc_new_address
- Closed
- links to
-
Review openjdk/jdk/17442