On Sat, 13 Feb 2021 07:10:03 GMT, Chris Plummer <[email protected]> wrote:
>>> > Yes, It is not known at present, but the heapInspect and heapDumper
>>> > accept an outputStream, and the actual parallel thread number can be got
>>> > by gang->active_workers() after WithUpdatedActiveWorkers. Then maybe some
>>> > code can be added there to record the actual number and finally print to
>>> > outputStream at the end.
>>>
>>> So you are proposing having the attach "heapdumpext" and "inspectheap"
>>> commands always print the number of threads used? I'm not so sure that is a
>>> good idea since there are probably existing tools and tests that parse the
>>> output, and changing its format could break them.
>>
>> Correct， if there is parsing tools (which I think there alreay are.), they
>> will be broken.
>> So I am afraid user can only tuning the thread number by expriments and some
>> common sense such as the core number of their server or the multi-processing
>> scenario.
>
>> Correct， if there is parsing tools (which I think there alreay are.), they
>> will be broken.
>> So I am afraid user can only tuning the thread number by expriments and some
>> common sense such as the core number of their server or the multi-processing
>> scenario.
>
> I'm ok with that. I think the other options we've explored are more trouble
> than they are worth.
Dear All,
May I say that we all agreed that "noparallel" is not necessary at present?
I think the PR 2519 and related CSR and issue could be all closed.
-------------
PR: https://git.openjdk.java.net/jdk/pull/2379
