[8u] RFR (S) 8190974: Parallel stream execution within a custom ForkJoinPool should obey the parallelism

[Tagir Valeev]

Hello!

Looks good to me. I don't think there are any hidden problems with this
fix. It was pretty straightforward and should not rely on anything which
might not be available in 8.

With best regards,
Tagir Valeev.

ср, 22 мая 2019 г., 22:47 Martin Buchholz <martinrb at google.com>:

> Too bad Paul Sandoz is no longer working on openjdk.
> Maybe Doug can review - he was involved with design of stream parallelism
> IIRC.
> I took a quick look - it Looks Good To Me.
>
> (But I have misgivings about any sort of auto-parallelization, including
> via parallel streams, which will become increasingly important as cpu
> counts increase)
>
> On Thu, May 2, 2019 at 10:45 AM Aleksey Shipilev <shade at redhat.com> wrote:
>
>> Original fix:
>>   https://bugs.openjdk.java.net/browse/JDK-8190974
>>   http://hg.openjdk.java.net/jdk/jdk/rev/e0041b182e31
>>
>> This is one of the things that was mentioned multiple times as the
>> pitfall in 8u. The patch itself
>> is simple enough to consider backporting and fixing that pothole. Once
>> related JDK-8154387 [1]
>> backport is in, this patch applies with reshuffling to 8u, except for the
>> test, which requires
>> fitting to the old shape of Stream tests.
>>
>> 8u webrev:
>>   http://cr.openjdk.java.net/~shade/8190974/webrev.8u.01/
>>
>> Testing: jdk_core tests, including the new test
>>
>> Thanks,
>> -Aleksey
>>
>> [1] https://bugs.openjdk.java.net/browse/JDK-8154387
>>
>>