From: Noah Goldstein <goldstein.w.n@gmail.com>
To: Wilco Dijkstra <Wilco.Dijkstra@arm.com>
Cc: GNU C Library <libc-alpha@sourceware.org>,
Adhemerval Zanella <adhemerval.zanella@linaro.org>,
"H.J. Lu" <hjl.tools@gmail.com>
Subject: Re: [PATCH v2] x86-64: Optimize bzero
Date: Thu, 24 Feb 2022 16:58:40 -0600 [thread overview]
Message-ID: <CAFUsyfKjzNw_nda31yn2FHRqqSAoutxcaTwdNVEA1EVs8F8B8g@mail.gmail.com> (raw)
In-Reply-To: <AS8PR08MB6534F65DF1686E29E8939B87833C9@AS8PR08MB6534.eurprd08.prod.outlook.com>
On Thu, Feb 24, 2022 at 7:16 AM Wilco Dijkstra <Wilco.Dijkstra@arm.com> wrote:
>
> Hi Noah,
>
> > The performance improvement for the build/link step for varying amount of
> > small functions per file was consistently in the ~.8% range. Not mind blowing
> > but I believe its a genuine improvement.
>
> I don't see how it's possible to get anywhere near 0.8%. I tried compiling a file with
> 10000 empty functions, and the latest __memset_exex_unaligned_erms takes about
> 1.16% of total time.
Smart sanity check I'll start using.
What method are you using for getting total function call overhead?
Using `perf record` + `pref report` and see a fair amount of variance but much
higher memset overheader (counting `_*unaligned_erms` and `*_unaligned`)
in `cc1` and `as`.
From average of 3 runs compiling file with 1/10/100/1000 functions I get:
1: 4.04%
10: 3.94%
100: 2.86%
1000: 2.68%
So its slightly less insane, arguing for the following speedups:
1: ~15%
10: ~15%
100: ~25% <--- this makes little to no sense
1000: ~15%
personally agree with you that those numbers seem to high though.
In the best case micro-benchmark that stressed the p5 bottleneck
this is about what we see.
>
> There are 81.5 million calls to memset in 48 billion cycles for this benchmark. That
> means 6.8 cycles per memset call on average. A 0.8% speedup would require making
> each memset 4.7 cycles faster, and that's not possible with bzero.
>
> To verify whether vpbroadcastb is a bottleneck I repeated it 16 times. This increased
> the memset percentage to 1.86%, however the total cycles didn't change measurably.
>
> I'm not sure how you're measuring this, but it's clear what you're seeing is not a
> speedup from bzero.
>
> Cheers,
> Wilco
next prev parent reply other threads:[~2022-02-24 22:58 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-15 13:38 Wilco Dijkstra
2022-02-23 8:12 ` Noah Goldstein
2022-02-23 12:09 ` Adhemerval Zanella
2022-02-24 13:16 ` Wilco Dijkstra
2022-02-24 15:48 ` H.J. Lu
2022-02-24 22:58 ` Noah Goldstein [this message]
2022-02-24 23:21 ` Noah Goldstein
2022-02-25 17:37 ` Noah Goldstein
2022-02-25 13:51 ` Wilco Dijkstra
2022-02-25 17:35 ` Noah Goldstein
-- strict thread matches above, loose matches on Subject: below --
2022-02-08 22:43 H.J. Lu
2022-02-08 23:56 ` Noah Goldstein
2022-02-09 11:41 ` Adhemerval Zanella
2022-02-09 22:14 ` Noah Goldstein
2022-02-10 12:35 ` Adhemerval Zanella
2022-02-10 13:01 ` Wilco Dijkstra
2022-02-10 13:10 ` Adhemerval Zanella
2022-02-10 13:16 ` Adhemerval Zanella
2022-02-10 13:17 ` Wilco Dijkstra
2022-02-10 13:22 ` Adhemerval Zanella
2022-02-10 17:50 ` Alejandro Colomar (man-pages)
2022-02-10 19:19 ` Wilco Dijkstra
2022-02-10 20:27 ` Alejandro Colomar (man-pages)
2022-02-10 20:42 ` Adhemerval Zanella
2022-02-10 21:07 ` Patrick McGehearty
2022-02-11 13:01 ` Adhemerval Zanella
2022-02-12 23:46 ` Noah Goldstein
2022-02-14 12:07 ` Adhemerval Zanella
2022-02-14 12:41 ` Noah Goldstein
2022-02-14 14:07 ` Adhemerval Zanella
2022-02-14 15:03 ` H.J. Lu
2022-05-04 6:35 ` Sunil Pandey
2022-05-04 12:52 ` Adhemerval Zanella
2022-05-04 14:50 ` H.J. Lu
2022-05-04 14:54 ` Adhemerval Zanella
2022-02-10 22:00 ` Alejandro Colomar (man-pages)
2022-02-10 19:42 ` Adhemerval Zanella
2022-02-10 18:28 ` Noah Goldstein
2022-02-10 18:35 ` 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=CAFUsyfKjzNw_nda31yn2FHRqqSAoutxcaTwdNVEA1EVs8F8B8g@mail.gmail.com \
--to=goldstein.w.n@gmail.com \
--cc=Wilco.Dijkstra@arm.com \
--cc=adhemerval.zanella@linaro.org \
--cc=hjl.tools@gmail.com \
--cc=libc-alpha@sourceware.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 read-only IMAP folder(s) and NNTP newsgroup(s).