[lug] Various Arch/Compiler Binaries living together

[Sean Reifschneider]

On Wed, Jun 26, 2002 at 12:15:50PM -0600, Ed Hill wrote:
>wouldn't be so surprised.  The P4 has *rotten* legacy x87 support.  The
>only way you get good floating point is by having a compiler that can

Intel CPUs have *ALWAYS* had pretty crappy floating-point.  Back in around
1995 I did some comparisons between a 60MHz HPPA machine (712/60) and a
Pentium 100MHz.  My tests were running a raytracer in integer and floating
point modes on both machines.

In integer mode, the HP machine was about 3x slower than the PC.  However
when running the same test using floating point on both machines, the HP
was easily able to keep up with the PC.  Remember, the HP is a RISC CPU, so
it's not suprising that 60MHz would be around 3x slower than the 100MHz
Pentium.

At the time, the Pentium was a state of the art box, and the HP was around
2 years old.  At the time of purchase, the HP was a step down from the top
end machine (which had an 80MHz CPU).

Sean
-- 
 "The guy with the switch-blade gets to deal."
 "Yeah, but that's not a switch-blade."  <Schick>  "I stand corrected."
Sean Reifschneider, Inimitably Superfluous <jafo at tummy.com>
tummy.com - Linux Consulting since 1995. Qmail, KRUD, Firewalls, Python