On Tue, 17 Feb 2004 16:22:25 +1300
Jim Cheetham <[EMAIL PROTECTED]> wrote:

> On Tue, 2004-02-17 at 15:57, Matthew Gregan wrote:
> > The performance difference between GCC optimising for i386 and i586 is
> > unlikely to be noticable for most tasks.
> 
> Indeed, and within the life-time of a specific version of a command,
> what is the likely comparison between the full configure/compile cycle,
> and the number of executions * optimisation speedup?
> 

I don't really do it for the speed optimisation (although i do use the
CFLAGS relevant to my cpu), but more for the escape from dependency hell,
and the USE flags feature, and the escape from upgrade hell (I never did
achive a smooth transition from one redhat/mandrake/debian version to
another.) With gentoo there is never such an upgrade, just the
transition to the next version of each package.

My choice, others have legit reasons for their choices.


> I suspect that a frequently-updating compile-from-source gentoo user is
> spending more CPU time per application than is healthy ... 
> 
> Especially with distributed.net, seti, protein folding, and other worthy
> causes clamouring for your CPU cycles ...
> 
> -jim

-- 
Nick Rout <[EMAIL PROTECTED]>

Reply via email to