-
Enhancement
-
Resolution: Fixed
-
P3
-
9
-
b16
Currently, when a mutator helps refining a dirty card buffer, it refines the current buffer it just filled.
In many situations this is not a good idea, because often the mutator will keep accessing the same memory areas again, immediately generating new card entries.
Change the algorithm to use the last (oldest) buffer.
There is a preliminary patch attached, needs perf testing.
In many situations this is not a good idea, because often the mutator will keep accessing the same memory areas again, immediately generating new card entries.
Change the algorithm to use the last (oldest) buffer.
There is a preliminary patch attached, needs perf testing.
- relates to
-
JDK-8137022 Concurrent refinement thread adjustment and (de-)activation suboptimal
-
- Resolved
-
-
JDK-8221360 Eliminate Shared_DirtyCardQ_lock
-
- Resolved
-