RFR: 8256265: G1: Improve parallelism in regions that failed evacuation [v2]

[Thomas Schatzl]

On Tue, 13 Sep 2022 10:10:04 GMT, Stefan Johansson <sjohanss at openjdk.org> wrote:

>> Thomas Schatzl has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 12 commits:
>> 
>>  - Merge branch 'master' of gh:tschatzl/jdk into 8256265-parallel-evac-failure
>>  - disable G1EvacuationFailureALot by default again
>>  - Remove unneeded clear metadata phase
>>  - Remove debug code
>>  - More refactoring
>>  - Initial cleanup
>>  - some refactoring, fix clearing of opt index in cset
>>  - fix test
>>  - some cleanup
>>  - Cleanup, phase names, fixes
>>  - ... and 2 more: https://git.openjdk.org/jdk/compare/37df5f56...2628451d
>
> src/hotspot/share/gc/g1/g1GCPhaseTimes.cpp line 106:
> 
>> 104:   _gc_par_phases[MergePSS] = new WorkerDataArray<double>("MergePSS", "Merge Per-Thread State (ms):", max_gc_threads);
>> 105:   _gc_par_phases[RestoreRetainedRegions] = new WorkerDataArray<double>("RestoreRetainedRegions", "Restore Retained Regions (ms):", max_gc_threads);
>> 106:   _gc_par_phases[RemoveSelfForwardsInChunks] = new WorkerDataArray<double>("RemoveSelfForwardsInChunks", "Remove Self Forwards In Chunks (ms):", max_gc_threads);
> 
> Is the "In Chunks" part really interesting to users? We do mention chunks below so the info is there. 
> 
> And if we could come up with a more user friendly name than "Remove Self Forwards" that would probably be good as well. Some thing like "Clean/Fix Failed Regions" maybe.

I can see your point about the "InChunks" postfix; however I would prefer to keep the phase meaningful wrt to what it does. Neither "Remove Self Forwards" or "Fix Failed Regions" helps end users in a big way, but the former helps at least us understand at a glance what this phase does without needing to translate "Fix Failed Regions" to "this phase removes the self forwards".

-------------

PR: https://git.openjdk.org/jdk/pull/9980