>
> On Mon, 28 Feb 2000, JF Martinez wrote:
>
> > If RedHat pushes towards resource hungry User interfaces like
> > Gnome/KDE and installations in Python who are looooooong (at least the
> > update was) on a Cyrix 686 at 133 Mhz (rated like a 166+) it seems
> > there is little reason to continue using the -m486 flag for compiling
> > since 486s seem far too underpowered for running (and installing)
> > RedHat 6.2.
>
> I'm running 6.2 without any problems on a 486. Text mode only, running my
> mail and a webserver with ~2500 hits/day.
>
That means one hit every 30 seconds. Several years ago Linux Journal
was getting 40K hits/day on a 486/33. But running mail on text mode
is not what I would call real use.
Also there is the time to install/upgrade. When upgreading from6.1 to
RH 6.2 beta it spent about ten minutes in the 'Preparing to install'
screen and this was on a Cyrix 166+ who is supposed to have about
three times the horsepower of a 486 DX2 66. That would have meant 30
minutes on a 486. Quite simply unbearable.
If you want to keep RedHat usable by 486s you have to keep a low
weight GUI and rewrite in C some parts of the install.
> > The only reason for keeping -m486 would be if the compiler would
> > generate buggy code with -mpentium.
>
> As long as we're using egcs 1.1.2, -mpentium isn't much of a gain. Even
> with -march=pentium, it's usually not noticable.
> gcc 2.95.x does a better job; gcc 2.96 will be really good with
> optimizations. At that point, changing the compiler flags will make more
> sense.
>
egcs -mpentium is nowadays just a marketing trick: it brings enough
improvements for looking good in benchmarks but the user will not even
notice them.
--
Jean Francois Martinez
Project Independence: Linux for the Masses
http://www.independence.seul.org
--
To unsubscribe:
mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
