bin/147451: "gcc -g -O2 -march=nocona" loops compiling devel/qt4-script

[Peter Jeremy]

The following reply was made to PR gnu/147451; it has been noted by GNATS.

From: Peter Jeremy <peterjeremy at acm.org>
To: Garrett Cooper <yanefbsd at gmail.com>
Cc: FreeBSD-gnats-submit at freebsd.org
Subject: Re: bin/147451: "gcc -g -O2 -march=nocona" loops compiling
 devel/qt4-script
Date: Sat, 5 Jun 2010 08:52:49 +1000

 --s2ZSL+KKDSLx8OML
 Content-Type: text/plain; charset=us-ascii
 Content-Disposition: inline
 Content-Transfer-Encoding: quoted-printable
 
 On 2010-Jun-03 18:49:35 -0700, Garrett Cooper <yanefbsd at gmail.com> wrote:
 >Worked for me on this machine, but I have over 12GB of RAM... so not
 >really a fair comparison.
 
 I tried again on a different host (2.5GHz Athlon 4850e with 2GB RAM
 and 9GB swap) and "-O2 -g -march=3Dnocona" ran out of swap after 7.6 hrs
 CPU.  "-O2", "-O2 -g" and "-O2 -march=3Dnocona" each ran in 25-30 secs.
 Interestingly, the working set remained fairly small (<1GB) so the
 system wasn't thrashing (though ZFS was very unhappy).
 
 >This is probably one of those ridiculous optimizer bugs like what
 >happened with x11-servers/xorg-server back in the day when 4.2.1 was
 >initially imported...
 
 It's probably a "well, don't do that then" case but if someone feels
 like chasing it down, I won't object.
 
 --=20
 Peter Jeremy
 
 --s2ZSL+KKDSLx8OML
 Content-Type: application/pgp-signature
 Content-Disposition: inline
 
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.14 (FreeBSD)
 
 iEYEARECAAYFAkwJg8EACgkQ/opHv/APuIfe+wCcCHiYmN4Rb1MKGj+uCIoBiclf
 0FsAoJLXxdWvaCc2k//QdgtHohJFfpAc
 =iVwR
 -----END PGP SIGNATURE-----
 
 --s2ZSL+KKDSLx8OML--