try lmbench, as well. 


On Wed, 24 Jan 2001, NetBSD Bob wrote:

> > I'd be interested to see how a kernel compiled with -Dinline= works for
> > you.  On a small-memory system, it might run quite a bit faster.
> > 
> > If you do that, you probably want to add -fno-inline-functions to the
> > compiler command line, too.
> 
> I am not able to drag in an MVII crate into the office to get it online
> for some test, but, I did manage to waltz in the door with a MV3100/10
> under my arm and not draw too many stares.  It has 16mb of ram and 5
> of the RZ25 drives, plus I have a couple of 1 gig things I can hang
> off the back.  If I load up one drive with various systems, what
> would be the ones of choice for the gurus to have loaded, and what
> kinds of compile switches, kernel configs, etc., do we need to get
> the most useful info out of the critter.  The is the slowest box
> I have, except for the MVII critters.   I will try rolling Bonnie,
> ttcp, and bytebench on it.  Is there anything else the gurus need
> to help figure out/obtain/etc., the needed info to push the slow
> VAXen along?
> 
> I could put up a generic account such as netbsd (original, right?)
> with a generic password, and give whomever wants su on it, and let
> them wrangle it out to their hearts content.  That way, the gurus
> could run up whatever they think is useful.

i woudlnt mind checking it out.. ill run my personal 'benchmark' , just a
number sieve.. but its my own bogus metric of machine speed :) 

btw i will probably be the proud owner of a VAX 4000/600 soon, just making
shipping arrangements for it now.. it's not supported yet, is it? if not,
does anybody have any docs? i dont think it woudl be very different from 
the vax4400 or 4500.. 

btw it turns out that i was mistaken about the speed of the pcc-generated 
code- it was in line with what ragge's post said, i had done my stuff on 
a ka650 and had miscalculated the speed ration from the 8600 to the 3600 
and gotten delusions of grandeur and speed from it.. :) 

happy hacking,
isildur