This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc compile-time performance
- From: Tim Prince <tprince at computer dot org>
- To: dewar at gnat dot com (Robert Dewar), dberlin at dberlin dot org,shebs at apple dot com
- Cc: ak at suse dot de, 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 18:52:11 -0700
- Subject: Re: gcc compile-time performance
- References: <20020518001420.36B7DF28C4@nile.gnat.com>
- Reply-to: tprince at computer dot org
On Friday 17 May 2002 17:14, Robert Dewar wrote:
> My own feeling here is that compile time performance is less important
> than run time performance. We have never seen a customer pushed in the
> direction of a proprietary compiler by compile time performance, but we
> have seen many concerns about runtime performance being slower.
I've seen the argument a few times lately "I like this compiler because it
has better run-time performance than MSVC, but it ought to be able to
accomplish that without taking longer than a non-optimized MSVC build." On a
bootstrap build, we're really pushing things to expect the first stage
compiler, built with no optimization, to build a complete set of compilers
for all languages with full optimization, and then asking for it to be fast
as well. It doesn't require -O2 to get as much optimization as gcc used to
have, but I don't see anyone willing to cut back even to -Os to economize on
bootstrap time.
--
Tim Prince