RFR: 8306738: Select num workers for safepoint ParallelCleanupTask [v2]
Coleen Phillimore
coleenp at openjdk.org
Thu Apr 27 19:52:58 UTC 2023
On Tue, 25 Apr 2023 14:34:06 GMT, Axel Boldt-Christmas <aboldtch at openjdk.org> wrote:
>> Select num workers for safepoint ParallelCleanupTask to be at most the number of parallel cleanup tasks `SAFEPOINT_CLEANUP_NUM_TASKS`
>>
>> Unsure if `WorkerThreads::max_workers()` is more correct than `WorkerThreads::created_workers()` or `WorkerThreads::active_workers()`.
>>
>> Tested tier1-3 Oracle platforms.
>
> Axel Boldt-Christmas has updated the pull request incrementally with one additional commit since the last revision:
>
> Use active_workers
I think I put the safepoint cleanup ordering code in assuming that the symbol table and string table rehashing takes a long time when it needs rehashing. But most/all of the time, it only notifies the ServiceThread that these tables need resizing. You can put those two at the end. From what I can tell from this, you only need 3 threads maximum to do the cleanup tasks. Aleksey's expected_workers() function makes a lot of sense to me.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/13616#issuecomment-1526265467
More information about the hotspot-runtime-dev
mailing list