RFR 8204947: Port ShenandoahTaskTerminator to mainline and make it default

[Wessam Hassanein]

Hi All,

We are glade to see this work and can provide our implementation which is
used in the paper for your reference. The work was initially created for
the CMS GC and then ported to G1. We have recently forward ported our work
to JDK11. We have not compared codewise to this work but feel free to do so
and use any/all you see fit.  I have asked Man to generate a webrev and
post it here.

Thanks,

Wessam Hassanein

On Fri, Dec 7, 2018 at 6:02 AM <zgu at redhat.com> wrote:

> Thanks, Thomas and Roman.
>
> -Zhengyu
>
> On Fri, 2018-12-07 at 14:58 +0100, Roman Kennke wrote:
> > > > On Thu, 2018-12-06 at 08:56 -0500, zgu at redhat.com wrote:
> > > > > > Looks good otherwise. I do not need a re-review for above
> > > > > > change.
> > > > > > I
> > > > > > will have this final change run through our testing one more
> > > > > > time
> > > > > > and will report back (tomorrow).
> > > > > >
> > > > >
> > > > > Awesome! In the meanwhile, I will run the latest through
> > > > > submit.
> > > >
> > > > Hs-tier1-7,jdk-tier1-3 of the final patch look good.
> > > >
> > > > One final note: could you update the reference to Google's paper
> > > > to
> > > > also show authors and other publication info?
> > > >
> > > > Something like the ACM ref on the ACM page so that people
> > > > preferring
> > > > paper can more easily find it, e.g.
> > > >
> > > > "Wessam Hassanein. 2016. Understanding and improving JVM GC work
> > > > stealing at the data center scale. In Proceedings of the 2016 ACM
> > > > SIGPLAN International Symposium on Memory Management (ISMM 2016).
> > > > ACM,
> > > > New York, NY, USA, 46-54. DOI:
> > > > https://doi.org/10.1145/2926697.2926706"
> > > >
> > > > No need for a re-review for me if you change it.
> > >
> > > Certainly, I will make the change as you suggested.
> > >
> > >
> > > Also, the latest submit passed. Now, I need a second reviewer and
> > > CSR
> > > approval to push, right?
> >
> > Patch looks good. Please wait for CSR approval.
> >
> > Thanks,
> > Roman
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20181207/bae83499/attachment.htm>