This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc compile-time performance
- From: Stan Shebs <shebs at apple dot com>
- To: Robert Dewar <dewar at gnat dot com>
- Cc: davem at redhat dot com, ak at suse dot de, dberlin at dberlin dot org, dhazeghi at pacbell dot net, gcc at gcc dot gnu dot org, neil at daikokuya dot demon dot co dot uk
- Date: Fri, 17 May 2002 18:31:38 -0700
- Subject: Re: gcc compile-time performance
- References: <20020518005526.1C904F28D2@nile.gnat.com>
Robert Dewar wrote:
>
> > I've seen quite the opposite, where compile time performance would
> > have made or broken someone's product ever shipping because bootstraps
> > took on the order of days. Programmers don't get paid to fart in
> > their chairs waiting for builds to finish so then can begin running
> > regression tests (well, actually, let's be honest, some do :-).
>
> Well I find the notion of a bootstrap taking days to be weird. I do a
> complete three cycle bootstrap of GNAT in about ten minutes on my notebook
> which is a slow machine (compared to say a $900 desktop). It is true that
> compared to fast compilers (e.g. Realia COBOL compiles at something like
> a million lines a minute on a fast PC), GCC is slow, but on today's fast
> machines it seems adequately fast to me.
I've noticed that people who've only ever worked with GCC don't
think anything is wrong, they just assume that's how fast
compilers are. People who come to GCC from CodeWarrior climb
the walls in frustration, because they don't see why their
2-minute compiles now have to take an hour.
Stan