This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Linking speed with large amount of debugging info (IO speed)
- From: Andrew_Pinski at PlayStation dot Sony dot Com
- To: Jakub Jelinek <jakub at redhat dot com>
- Cc: binutils at sourceware dot org
- Date: Wed, 13 Jun 2007 12:00:27 -0700
- Subject: Re: Linking speed with large amount of debugging info (IO speed)
Jakub Jelinek <jakub@redhat.com> wrote on 06/12/2007 02:06:24 AM:
> On Mon, Jun 11, 2007 at 02:35:47PM -0700,
Andrew_Pinski@PlayStation.Sony.Com wrote:
> > The difference between linking with debugging and without is 3 minutes
vs
> > 30 seconds which seems a bit high.
>
> See http://sources.redhat.com/bugzilla/show_bug.cgi?id=3111
> - the most likely culprit.
I tried the top of the trunk as of yeasterday and I still have a huge
speed issue.
Here is the timings I got:
apinski@debian:~/hs$
~/src/gcc-sony-4.1/sony41clean1/test/install/cell/host-linux/ppu/bin/ppu-lv2-ld
--version
GNU ld (GNU Binutils) 2.17.50.20070612
Copyright 2007 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License. This program has absolutely no warranty.
apinski@debian:~/hs$ time
~/src/gcc-sony-4.1/sony41clean1/test/install/cell/host-linux/ppu/bin/ppu-lv2-gcc
@link.rsp
real 2m26.355s
user 0m24.734s
sys 0m6.184s
Notice how the time is not spent in user (or system) time but only in real
time.
I have nothing else running on this machine except some terminals and the
load average is 0.0.
2.17's result was:
real 2m43.739s
user 0m28.542s
sys 0m6.088s
So it is faster than 2.17 but still slow.
Thanks,
Andrew Pinski