Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 11 Jan 2009 07:23:06 -0800 (PST)
From:      "Pedro F. Giffuni" <giffunip@tutopia.com>
To:        "O. Hartmann" <ohartman@mail.zedat.fu-berlin.de>
Cc:        freebsd-current@freebsd.org
Subject:   Re: Alternatives to gcc (was Re: gcc 4.3: when will it become standard compiler?)
Message-ID:  <458984.49823.qm@web32708.mail.mud.yahoo.com>
References:  <61484.71762.qm@web32708.mail.mud.yahoo.com> <20090111044448.GC5661@server.vk2pj.dyndns.org> <4969CC6D.6030707@mail.zedat.fu-berlin.de>

next in thread | previous in thread | raw e-mail | index | archive | help
=0A=0A----- Original Message ----=0A...=0A> =0A> Well, initially my questio=
n was triggered by reading a performance duell=0A> between FreeBSD 7/8, mos=
t recent U(n)buntu and OpenSolaris and someone=0A> stated the 3% performanc=
e gain of U(n)buntu over FreeBSD was due to the=0A> gcc4.3 compiler, which =
generates more efficient code. 3% mean=0A> performance gain could mean (as =
I made this experience) a better=0A> advantage in some special cases and ha=
ving in mind numerical modelling=0A> running on my lab's FreeBSd box (yet, =
but I think this is about to=0A> change and move towards a RH Linux system =
due to the better support of=0A> HPC and, a pitty, our admins build the clu=
ster with RH and not FBSD).=0A> =0A=0AEven when it can be measured, perform=
ance can be very subjective, performance=0Adepends on many factors: the thr=
eading libraries, the options used to build the =0Apackages, the filesystem=
s and maybe even the position of the moon ;-). Most of =0Amy numerical pack=
ages don't depend on the system compiler but rather depend on =0Awhat the=
=A0ports system=A0uses=A0as the=A0Fortran compiler so you will be glad to k=
now =0Athat we are indeed using gcc4.3 since last week.=0A=0A> =0A> Well, a=
s I understand the discussion about the binutils (there seems to=0A> exist =
a very similar problemacy), did RH already cut off the leashes by=0A> intro=
ducing their elftools? Correct me, if I'm wrong.=0A> =0A=0AWe already have =
our own libelf and related utilities however the tough part seems =0Ato be =
having a good assembler that supports all our platforms. I understand the R=
H =0Aelftools have that but I don't know their current state.=A0Also the=A0=
maintainers of these =0Autilities are known to be rather unfriendly with ot=
her camps.=0A=0APedro.=0A=0A=0A      



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?458984.49823.qm>