peter.wood at worldonline.dk peter.wood at worldonline.dk
Tue Sep 5 13:28:24 PDT 2000

On Tue, Sep 05, 2000 at 08:49:53AM -0400, Gerard Beekmans wrote:
> On Mon, 04 Sep 2000, you wrote:
> > The main problem I have is that I am not familiar with exactly *what* may
> > be causing the problem.  I was more or less asking if there are any
> > packages that anyone KNOWS offhand doesn't work at such high optimisations.
> >  I am really wanting everything possible to be optimised.  Is it possible
> > to use optimisations when compiling glibc or should I compile it with the
> > default -O2 (maybe -O3)
> In addition to what others have said: it's not possible to tell what breaks 
> and what doesn't. David mentioned him having problems with shadow password 
> suite, I don't. Others have mentioned binutils and glibc problems, I don't. 
> It's all very machine dependant. Till now I've compiled each and every 
> package in the LFS book with -O3 -march=i686 -mcpu=i686 (and on shadowfax 
> with -i586) and everything is working just dandy. Although very very 
> ocasionally things don't work. Like KDE2-Beta4 - ksplash crashes wtih a 
> segmentation fault whereas it doesn't do that on other systems.
> No, the reason that optimization works seemingly well here doesn't mean I 
> have some super computer. It's the sum of the hardware that counts. Perhaps 
> I'm lucky to have a stable CPU (though it's not that new or anything - intel 
> celeron 533) or whatever else is caused. Perhaps it's because i'm the LFS 
> project leader and software likes me a lot for that reason, who knows.

I compiled everything (as far as I remember, ha, ha) with "-O3 -mcpu=i586
-march=i586", on an AMD K6 233Mhz and everything works (so far).  And my answer
to the question "Why go to the trouble of bothering with optimisation?" is
"Because I can."  Just like some car freak might tune his vehicle endlessly,
for insignificant performance gains, or maybe even performance losses!  I like
fiddling with my software (sounds disgusting doesn't it?) so I do.


More information about the lfs-dev mailing list