[11u] RFR (S) 8214922: Aarch64: Add vectorization support for fmin/fmax

Langer, Christoph christoph.langer at sap.com
Wed Mar 10 10:57:39 UTC 2021


Hi,

I agree, the format changes from 8241911 should be added to this backport.

Best regards
Christoph

> -----Original Message-----
> From: Volker Simonis <volker.simonis at gmail.com>
> Sent: Montag, 8. März 2021 13:15
> To: Dmitry Chuyko <dmitry.chuyko at bell-sw.com>
> Cc: jdk-updates-dev at openjdk.java.net; aarch64-port-
> dev at openjdk.java.net; Langer, Christoph <christoph.langer at sap.com>
> Subject: Re: [11u] RFR (S) 8214922: Aarch64: Add vectorization support for
> fmin/fmax
> 
> Hi Dmitry,
> 
> in general, the downport looks good to me.
> 
> I only wonder if we shouldn't integrate the format changes from
> 8241911 which were skipped in December when 8241911 was downported
> [1,2] because at that time this change wasn't in 11u. I'm fine either
> way and leave the final decision to the 11u maintainers.
> 
> Thank you and best regards,
> Volker
> 
> [1] https://mail.openjdk.java.net/pipermail/jdk-updates-dev/2020-
> December/004474.html
> [2] https://mail.openjdk.java.net/pipermail/jdk-updates-dev/2020-
> December/004498.html
> 
> On Thu, Feb 18, 2021 at 9:27 AM Dmitry Chuyko <dmitry.chuyko at bell-
> sw.com> wrote:
> >
> > Hello,
> >
> > Original RFE: https://bugs.openjdk.java.net/browse/JDK-8214922
> >
> > Original patch applies almost cleanly except 2 added lines in
> > src/hotspot/share/opto/vectornode.cpp because of missing part of
> context
> > (JDK-8214751 not ported), they are added manually.
> >
> > 11u webrev:
> http://cr.openjdk.java.net/~dchuyko/8214922/webrev.11u.00/
> >
> > Testing: tier1, tier2, TestFpMinMaxIntrinsics.
> >
> > --
> > Thanks,
> > -Dmitry
> >


More information about the jdk-updates-dev mailing list