[11u] RFR 8245512: CRC32 optimization using AVX512 instructions

[Langer, Christoph]

Hi,

with the discussion/information so far I also tend to accept it. Andrew, please make the final call ��

Best regards
Christoph


> -----Original Message-----
> From: Hohensee, Paul <hohensee at amazon.com>
> Sent: Donnerstag, 7. Januar 2021 15:13
> To: Florian Weimer <fweimer at redhat.com>; Andrew Haley
> <aph at redhat.com>
> Cc: Doerr, Martin <martin.doerr at sap.com>; Viswanathan, Sandhya
> <sandhya.viswanathan at intel.com>; Langer, Christoph
> <christoph.langer at sap.com>; Lindenmaier, Goetz
> <goetz.lindenmaier at sap.com>; Vladimir Kozlov
> <vladimir.kozlov at oracle.com>; jdk-updates-dev at openjdk.java.net;
> hotspot-compiler-dev at openjdk.java.net
> Subject: Re: [11u] RFR 8245512: CRC32 optimization using AVX512 instructions
> 
> If AVX512 support were to be removed, the addition of this patch would not
> make removal more difficult. I agree with Martin: given that the patch is
> small, well tested, and of limited scope, I'm in favor of accepting it.
> 
> Thanks,
> Paul
> 
> -----Original Message-----
> From: hotspot-compiler-dev <hotspot-compiler-dev-
> retn at openjdk.java.net> on behalf of Florian Weimer
> <fweimer at redhat.com>
> Date: Thursday, January 7, 2021 at 4:25 AM
> To: Andrew Haley <aph at redhat.com>
> Cc: "Doerr, Martin" <martin.doerr at sap.com>, "Viswanathan, Sandhya"
> <sandhya.viswanathan at intel.com>, "Langer, Christoph"
> <christoph.langer at sap.com>, "Lindenmaier, Goetz"
> <goetz.lindenmaier at sap.com>, Vladimir Kozlov
> <vladimir.kozlov at oracle.com>, "jdk-updates-dev at openjdk.java.net" <jdk-
> updates-dev at openjdk.java.net>, "hotspot-compiler-
> dev at openjdk.java.net" <hotspot-compiler-dev at openjdk.java.net>
> Subject: RE: [11u] RFR 8245512: CRC32 optimization using AVX512 instructions
> 
> * Andrew Haley:
> 
> > We don't know the future of AVX-512. Rumours say it's not going
> > to be included in Alder Lake-S, Intel’s first 10nm mainstream
> > desktop processor. I don't think the 10900K Comet Lake supports
> > it either. But I admit it does look like Xeons are going to
> > support AVX-512 for some time, so perhaps.
> 
> AVX-512 support has started to appear in non-server CPUs with Tiger
> Lake:
> 
> <https://ark.intel.com/content/www/us/en/ark/products/208656/intel-
> core-i3-1125g4-processor-8m-cache-up-to-3-70-ghz.html>
> 
> There certainly were issues around power consumption and the internal
> turbo license for previous implementations of AVX-512.  But that's
> largely a question for mainline defaults: If the change is not generally
> beneficial and only looks good in certain micro-benchmarks, it's
> probably not appropriate for mainline either, at least not by default.
> I don't know to what extent it is necessary to reconsider mainline
> optimization choices in a backport.
> 
> Thanks,
> Florian
> --
> Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn,
> Commercial register: Amtsgericht Muenchen, HRB 153243,
> Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael
> O'Neill
>