From: Noah Goldstein <goldstein.w.n@gmail.com>
To: Eric Dumazet <edumazet@google.com>
Cc: tglx@linutronix.de, mingo@redhat.com,
Borislav Petkov <bp@alien8.de>,
dave.hansen@linux.intel.com, X86 ML <x86@kernel.org>,
hpa@zytor.com, peterz@infradead.org, alexanderduyck@fb.com,
open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v1] x86/lib: Optimize 8x loop and memory clobbers in csum_partial.c
Date: Fri, 26 Nov 2021 12:50:28 -0600 [thread overview]
Message-ID: <CAFUsyfJ33cKFQdUagHQ_b4N80CfBtGQZhyA4CN_JLgEmXEX=DA@mail.gmail.com> (raw)
In-Reply-To: <CANn89iJxMhGfp364rPu6p_ZLrKnM1qWF_NWrw4_oL_KG+piByg@mail.gmail.com>
On Fri, Nov 26, 2021 at 12:27 PM Eric Dumazet <edumazet@google.com> wrote:
>
> On Fri, Nov 26, 2021 at 10:17 AM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> >
>
> >
> > Makes sense. Although if you inline I think you definitely will want a more
> > conservative clobber than just "memory". Also I think with 40 you also will
> > get some value from two counters.
> >
> > Did you see the number/question I posted about two accumulators for 32
> > byte case?
> > Its a judgement call about latency vs throughput that I don't really have an
> > answer for.
> >
>
> The thing I do not know is if using more units would slow down the
> hyper thread ?
There are more uops in the two accumulator version so it could be concern
iff the other hyperthread is bottlenecked on p06 throughput. My general
understanding is this is not the common case and that the very premise of
hyperthreads is that most bottlenecks are related to memory fetch or resolving
control flow.
>
> Would using ADCX/ADOX would be better in this respect ?
What would code using those instructions look like? Having trouble
seeing how to use them here.
next prev parent reply other threads:[~2021-11-26 19:31 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-25 19:38 [PATCH v1] x86/lib: Optimize 8x loop and memory clobbers in csum_partial.c Noah Goldstein
2021-11-26 1:50 ` Eric Dumazet
2021-11-26 2:15 ` Noah Goldstein
2021-11-26 2:18 ` Noah Goldstein
2021-11-26 2:38 ` Noah Goldstein
2021-11-28 19:47 ` David Laight
2021-11-28 20:59 ` Noah Goldstein
2021-11-28 22:41 ` David Laight
2021-12-02 14:24 ` David Laight
2021-12-02 15:01 ` Eric Dumazet
2021-12-02 20:19 ` Noah Goldstein
2021-12-02 21:11 ` David Laight
2021-11-26 16:08 ` Eric Dumazet
2021-11-26 18:17 ` Noah Goldstein
2021-11-26 18:27 ` Eric Dumazet
2021-11-26 18:50 ` Noah Goldstein [this message]
2021-11-26 19:14 ` Noah Goldstein
2021-11-26 19:21 ` Eric Dumazet
2021-11-26 19:50 ` Noah Goldstein
2021-11-26 20:07 ` Eric Dumazet
2021-11-26 20:33 ` Noah Goldstein
2021-11-27 0:15 ` Eric Dumazet
2021-11-27 0:39 ` Noah Goldstein
2021-11-26 18:17 ` Eric Dumazet
2021-11-27 4:25 ` [PATCH v2] " Noah Goldstein
2021-11-27 6:03 ` Eric Dumazet
2021-11-27 6:38 ` Noah Goldstein
2021-11-27 6:39 ` [PATCH v3] " Noah Goldstein
2021-11-27 6:51 ` Eric Dumazet
2021-11-27 7:18 ` Noah Goldstein
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAFUsyfJ33cKFQdUagHQ_b4N80CfBtGQZhyA4CN_JLgEmXEX=DA@mail.gmail.com' \
--to=goldstein.w.n@gmail.com \
--cc=alexanderduyck@fb.com \
--cc=bp@alien8.de \
--cc=dave.hansen@linux.intel.com \
--cc=edumazet@google.com \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=tglx@linutronix.de \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).