michel beausejour (mbeausej@qc.bell.ca) wrote:
: I was making some comparisons between Netbsd and FreeBSD. I've compiled povray2.1
: on both platforms and tried the result programs on both platforms.
: 
: Under NetBSD i've compiled the program with gcc with options -m68040 and -m68881
: under FreeBSD i've used the options -m486.
: 
: NetBSD                                                    FreeBSD 2.0.5
What NetBSD?
: A3000 2meg chip/8megsFAST                             Dell 486Dx33 12megs (70ns)
: (8megs SCZIP 70ns)
: PicassoII
: A3640 (68040 25MHZ)

Was it -1.0 or an elderly current?

Until very recently, the sbic.c scsi hardware driver for the A3000
wasn't able to do relect and concurrent IO... meaning that any swapping
or i/o activity would stop anything else from happening at the same time.

Generally, swapping speed depends on the speed and number of disks
used... you get a speedup of 50% or so if you use a seperate disk for
swap than for /usr & data. What was your setup with the dell and the A3000?

And then, NetBSD + an color X server + a big application is too much for
8 Megs, if you want to avoid swap... I saw a definite speedup when going
from 8 to 12 Megs of fastmem. You might want to pull 4 Megs out of the
FreeBSD/86 machine and try again.

: also i've used a precompile version of povray (povray040 on aminet) to test
: under amigados

This one is optimized (AFAIremember) with a newer, better gcc than the
one shipped with NetBSD. NetBSD needs a stable, all-bugs-known compiler,
so the introduction of a newer gcc will be delayed at least until after the 
next OS release, I guess, which is scheduled for 95Nov17.

Regards,
	Ignatios Souvatzis