On Sat, 24 Jan 1998, David Brownlee wrote:
> On Fri, 23 Jan 1998, emanuel stiebler wrote:

> > P.S. the tps number i got from a performance chart from the dec site. I
> > think, that is more interresting than the pure CPU power.
> > 
> 	Indeed - but most vax people are more familiar with VUPs - ideally
> 	we should have both.

Quick question.  Aren't the tps values tightly coupled with the OS and
peripherals used?  Surely the DEC advertised tps value for say a
VAXstation 3100/M30 doesn't hold for the same machine running NetBSD/vax
with a PIO hacked kernel, running on a junky ST1480, which sounds like a
jet engine, that I scavenged from a NeXT :-) I'd imagine that the values
were computed for VMS, and probably with the standard issue peripherals
for the models in question.  This might explain why machines of equal VUPs
have different tps values, one could have been issued with a superior disk
drive that had a better transfer rate.  Or am I getting wrong the gist
of what goes into computing a system's tps? 

I still prefer the VXcf values, as they measure the novelty of one's VAX.  

-brian.
---
Brian "JARAI" Chase | http://world.std.com/~bdc/ | VAXZilla LIVES!!!