Let me preface this with pointing out that I own at least 2 PDP-11s, a VAX 
3100/40, a PC532 & a VME532, scads of old 68k VME stuff, a couple of (old) 
Alphas, an ICM-3216, some Sparc32s, numerous strange MIPS and ARM things, 
and a bunch of other fun, antiquated toys, most of which I've at least tried 
to run NetBSD on...oh, yeah...and I know I'm going to regret saying all of 
this... 

Dave McGuire writes: 

> On May 26, Ken Seefried wrote:
>> > now you've got me wondering if NetBSD runs on VAX 11/730.  0.3 VUPS.
>> > probably the slowest 32bit machine aside from the microvax I.  :)  
>> 
>> There is a Sun2 NetBSD port.  68010-based, not more than 12MHz (I would 
>> guess 8MHz), Multibus I.  I can't imagine it's faster that an 11/730 (a 
>> 386/16 is a relative speed deamon in comparison, even in the SX variety). 
> 
>   Sun2 systems run at 10MHz.  I ran both many years ago; the Sun2
> seemed quite a bit faster at most things.

Subjective, but fair enough.  I can say that my ISI 68010-based VME machine 
(4.2BSD) was quite a bit slower than the 11/780 (4.[23]BSD) I had access to, 
and that my 386/33 under Linux 1.0.x was faster, subjectively.  Arm wrestle 
for whose subjective view is right? 

>   My old VAX-11/730 would kick the crap out of pretty much any i386,
> though, when actual multitasking was involved.  

As apposed to that fake multitasking?  Anyone want to benchmark, say, a MV2 
(definately faster than an 11/730) and, say, a Compaq SystemPro/386?  I know 
where I would place my bet. 

I'll assume things like a Sequent Symmetry and NCR 3xxx-series are 
off-limits. 

> Don't underestimate
> the power of a well-designed architecture like the VAX.  Or, more to
> the point, don't OVERestimate the suckfulness of a horribly-designed
> architecture like x86!  

With all due respect (and I do indeed have great respect for Dave), I think 
the only thing that I underestimate is the devotion that people have to 
their favorite architecure, or their distaste for anything Intel, facts be 
damned. 

I should know better...I once said "Macs ain't so bad" at a Unix users group 
meeting.  I still get death threats...:-) 

> The only way the PeeCee weenies get any CPU
> performance at all out of those machines is by cranking them up to
> utterly ridiculous clock rates in compensation for a poorly designed
> processor. :)

Um...sure.  That old line. 

If I had a dime for every time I heard: 

"If <insert your favorite dead and buried CPU architecture> ran at as fast a 
clock as those drooling shuffle-tards over at Intel (who IMHO couldn't 
design a CPU to save their lives, and frankly can barely walk upright, and 
who molest goats, damn the billions and billions they make in the buisiness) 
ran their CPU, you'd see how much better <insert your favorite dead and 
buried CPU architecture> is, DAMMIT!  The <insert your favorite dead and 
buried CPU architecture> is obviously vastly faster (or just 'better', or 
'kicks the crap out of') anything that Intel produces.  Any benchmark that 
you name that disputes that is falsified, contrived, not-real-world, 
slave-to-the-vendor, or otherwise trivially discredited by the true 
disciples of <insert your favorite dead and buried CPU architecture>.  Only 
an idiot (that's you, all you millions and millions of people who are only 
pretending to actually do real work...which of course can't happen without 
the perfection displayed in <insert your favorite dead and buried CPU 
architecture>) don't realize this". 

I'm sure the buggy whip guys said something similar about Ford. 

BTW...you can substitute "CPU architecture" with "operating system" and 
"Intel" for "Microsoft" and get the same value. 

Sorry to bust your chops, Dave. 

Ken