This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: egcs -O6 fails to compile make 3.75 on i686-pc-linux-gnulibc1
- To: Daniel dot Veillard at imag dot fr, iskra at student dot uci dot agh dot edu dot pl
- Subject: Re: egcs -O6 fails to compile make 3.75 on i686-pc-linux-gnulibc1
- From: mrs at wrs dot com (Mike Stump)
- Date: Thu, 18 Sep 1997 09:38:08 -0700
- Cc: egcs at cygnus dot com
> Date: Thu, 18 Sep 1997 17:05:08 +0200 (MET DST)
> From: Kamil Iskra <kamil@dwd.interkom.pl>
> To: Daniel.Veillard@imag.fr
> cc: egcs@cygnus.com
> I noticed that some people use EGCS with optimisation level higher
> than 3. Does this make any sense?
Yes.
> According to the manual, using levels >=4 is equivalent to level 3.
The manual will be wrong. We know this. :-) This is why we use -O9,
memorize it, train our fingers to type it... While it may seem
pointless today, in the future, -O9 may run some really hairy
optimizations, and we want to use them... It we were trained to use
-O3, the day would come when -O4 did a little something more. Then
-O5.
Also, it is a sub-conscience wish on our part... We want the
generated code to be better than it is. :-) Kinda like praying before
you compile and test. While praying may in fact help, better to
engineer in the quality and robustness.
> I've heard that pgcc supports higher optimisation levels than 3, but
> is this also true for EGCS?
In time...