This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Compilation performance comparison of 3.5.0 and TreeSSA trees on MICO sources as requested in: [tree-ssa] Merge status 2004-05-03
- From: Jan Hubicka <hubicka at ucw dot cz>
- To: Laurent GUERBY <laurent at guerby dot net>
- Cc: Robert Dewar <dewar at gnat dot com>, Joe dot Buck at synopsys dot COM,dnovillo at redhat dot com, gcc at gcc dot gnu dot org, kenner at vlsi1 dot ultra dot nyu dot edu
- Date: Thu, 6 May 2004 01:13:10 +0200
- Subject: Re: Compilation performance comparison of 3.5.0 and TreeSSA trees on MICO sources as requested in: [tree-ssa] Merge status 2004-05-03
- References: <20040505174211.EEDB3F2A09@nile.gnat.com> <1083780566.25832.34.camel@pc>
> A -Og flag that would do all optimizations that
>
> - don't take too much time and memory ressources
> - still allow perfect debugability
> - tend to generate smalled code and debug information
>
> Would be just perfect :).
>
> I would fix a rule to allow up to 5-10%
> more compile time and memory at -Og than -O0
> and zero gdb breakage.
Actually I did some measurements for my GCC summit paper, and for most of C
sources I played around, we consume more peak memory at -O0 than on
-O1...
Just my 2 cents...
Honza
>
> But such optimization should be likely to reduce compile time anyway...
>
> Laurent
>