From: Matteo Frigo <athena@fftw.org>
To: Xavier Leroy <Xavier.Leroy@inria.fr>
Cc: Dmitry Bely <dmitry.bely@gmail.com>, Caml List <caml-list@inria.fr>
Subject: Re: [Caml-list] Ocamlopt x86-32 and SSE2
Date: Sun, 10 May 2009 19:12:49 -0400 [thread overview]
Message-ID: <87pregtuvi.fsf@fftw.org> (raw)
In-Reply-To: <4A0407A9.4000009@inria.fr> (Xavier Leroy's message of "Fri\, 08 May 2009 12\:21\:29 +0200")
Do you guys have any sort of empirical evidence that scalar SSE2 math is
faster than plain old x87?
I ask because every time I tried compiling FFTW with gcc -m32
-mfpmath=sse, the result has been invariably slower than the vanilla x87
compilation. (I am talking about scalar arithmetic here. FFTW also
supports SSE2 2-way vector arithmetic, which is of course faster.)
I also remember trying similar experiments with other numerical code in
the Pentium 4 dark ages, with similar results. I don't see any reason
why this should be the case, and maybe this is just a problem of gcc,
but I don't think you should automatically assume that SSE2 math is
faster without running a few experiments first.
Regards,
Matteo Frigo
next prev parent reply other threads:[~2009-05-10 23:13 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-04-28 19:36 Ocamlopt code generator question Dmitry Bely
[not found] ` <m27i13tofi.fsf@Pythagorion.local.i-did-not-set--mail-host-address--so-tickle-me>
2009-04-29 16:50 ` Dmitry Bely
2009-04-29 20:04 ` Jeffrey Scofield
2009-05-05 9:24 ` [Caml-list] " Xavier Leroy
2009-05-05 9:41 ` Dmitry Bely
2009-05-05 14:15 ` Jean-Marc Eber
2009-05-05 14:58 ` Sylvain Le Gall
2009-05-05 15:21 ` [Caml-list] " David Allsopp
2009-05-05 15:59 ` Dmitry Bely
[not found] ` <4A006410.8000205@lexifi.com>
2009-05-05 16:26 ` Dmitry Bely
2009-05-05 15:14 ` [Caml-list] " Jon Harrop
2009-05-08 10:21 ` [Caml-list] Ocamlopt x86-32 and SSE2 Xavier Leroy
2009-05-10 11:04 ` David MENTRE
2009-05-11 2:43 ` Jon Harrop
2009-05-11 3:43 ` Stefan Monnier
2009-05-11 5:38 ` [Caml-list] " Jon Harrop
2009-05-10 23:12 ` Matteo Frigo [this message]
2009-05-11 2:45 ` [Caml-list] " Jon Harrop
2009-05-11 7:55 ` Dmitry Bely
[not found] <20090509100004.353ADBC5C@yquem.inria.fr>
2009-05-09 11:38 ` CUOQ Pascal
2009-05-10 1:52 ` [Caml-list] " Goswin von Brederlow
2009-05-10 2:16 ` Seo Sanghyeon
2009-05-10 3:50 ` Jon Harrop
2009-05-11 8:05 ` Dmitry Bely
2009-05-11 9:26 ` Jon Harrop
2009-05-11 8:43 ` Dmitry Bely
2009-05-11 13:47 ` Jon Harrop
2009-05-11 9:12 ` Andrey Riabushenko
2009-05-10 8:56 ` CUOQ Pascal
2009-05-10 14:47 ` [Caml-list] " Richard Jones
2009-05-10 19:25 ` Florian Weimer
[not found] <20090511043120.976EBBC67@yquem.inria.fr>
2009-05-11 7:10 ` Pascal Cuoq
2009-05-12 9:37 ` [Caml-list] " Xavier Leroy
2009-05-12 12:40 ` Richard Jones
2009-05-13 22:30 ` Florian Weimer
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=87pregtuvi.fsf@fftw.org \
--to=athena@fftw.org \
--cc=Xavier.Leroy@inria.fr \
--cc=caml-list@inria.fr \
--cc=dmitry.bely@gmail.com \
/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).