From: Jonathan Wakely <jwakely.gcc@gmail.com>
To: Ico <gcc@zevv.nl>
Cc: gcc-help <gcc-help@gcc.gnu.org>
Subject: Re: Floating point performance issue
Date: Tue, 20 Dec 2011 10:34:00 -0000 [thread overview]
Message-ID: <CAH6eHdSz-g2QFJG3RRGCWvOr8azXy51XsjGpJHm-gvLVguYmFw@mail.gmail.com> (raw)
In-Reply-To: <1324376083-sup-8851@pruts.nl>
On 20 December 2011 10:20, Ico wrote:
>
> Still, I'm not sure if sse is part of the problem and/or solution.
It's the solution.
> I have been reducing the program to see what the smallest code is that still
> shows this behaviour. Latest version is below.
>
>
> $ gcc -msse -mfpmath=sse -O3 -march=native test.c
What is "native" for your system, i686? (also, what does gcc -dumpmachine show?)
i686 doesn't support SSE, you need at least pentium3.
Remove the -msse and see if you get a warning telling you SSE
instructions are disabled.
Try -march=pentium3 -mfpmath=sse instead (without -msse)
If you don't have at least a pentium3, you're stuck with the 387 FP
registers, and have to use horrible code.
> $ time ./a.out 0.9
>
> real 0m2.653s
> user 0m2.648s
> sys 0m0.002s
That looks as though you're still not using SSE registers.
> $ time ./a.out 0.001
>
> real 0m0.144s
> user 0m0.140s
> sys 0m0.002s
next prev parent reply other threads:[~2011-12-20 10:34 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-12-20 9:52 Ico
2011-12-20 10:05 ` Marcin Mirosław
2011-12-20 10:20 ` Ico
2011-12-20 10:34 ` Jonathan Wakely [this message]
2011-12-20 10:43 ` Ico
2011-12-20 11:24 ` Vincent Lefevre
2011-12-20 11:51 ` Dario Saccavino
2011-12-20 12:02 ` Ico
2011-12-20 12:12 ` Vincent Lefevre
2011-12-20 12:28 ` Tim Prince
2011-12-20 12:43 ` Segher Boessenkool
2011-12-20 13:02 ` Vincent Lefevre
2011-12-20 19:51 ` Segher Boessenkool
2011-12-20 21:02 ` Vincent Lefevre
2011-12-21 4:36 ` Segher Boessenkool
2011-12-21 6:15 ` Segher Boessenkool
2011-12-23 20:25 ` Vincent Lefevre
2011-12-20 13:43 ` David Brown
2011-12-20 13:58 ` Vincent Lefevre
2011-12-20 14:25 ` David Brown
2011-12-20 15:05 ` Vincent Lefevre
2011-12-20 15:44 ` David Brown
2011-12-20 16:18 ` Vincent Lefevre
2011-12-20 22:32 ` David Brown
2011-12-23 20:11 ` Vincent Lefevre
2011-12-24 7:38 ` Vincent Lefevre
2011-12-24 11:11 ` David Brown
2011-12-26 1:15 ` Vincent Lefevre
2011-12-26 11:48 ` David Brown
2011-12-26 13:07 ` Vincent Lefevre
2011-12-26 13:37 ` Tim Prince
2011-12-26 14:01 ` Vincent Lefevre
2011-12-26 14:39 ` Tim Prince
2011-12-26 16:40 ` Vincent Lefevre
2011-12-21 1:19 ` Miles Bader
2011-12-21 2:19 ` David Brown
2011-12-21 4:03 ` Miles Bader
2011-12-21 8:32 ` David Brown
2011-12-21 9:02 ` Miles Bader
2011-12-21 9:23 ` David Brown
2011-12-21 11:58 ` Miles Bader
2011-12-21 16:49 ` David Brown
2011-12-22 3:23 ` Miles Bader
2011-12-23 21:15 ` Vincent Lefevre
2011-12-24 17:18 ` David Brown
2011-12-26 8:12 ` Vincent Lefevre
2011-12-26 13:00 ` David Brown
2011-12-26 13:22 ` Vincent Lefevre
2011-12-20 15:45 ` Jeff Kenton
2011-12-20 11:44 ` David Brown
2011-12-20 11:49 ` David Brown
2011-12-20 10:46 ` Marc Glisse
2011-12-20 11:11 ` Ico
2011-12-20 11:16 ` Vincent Lefevre
2011-12-20 12:00 ` Vincent Lefevre
2011-12-20 12:21 ` Tim Prince
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=CAH6eHdSz-g2QFJG3RRGCWvOr8azXy51XsjGpJHm-gvLVguYmFw@mail.gmail.com \
--to=jwakely.gcc@gmail.com \
--cc=gcc-help@gcc.gnu.org \
--cc=gcc@zevv.nl \
/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).