[Info-vax] Why it is a good idea that OpenVMS isn't on x86-64 just yet

[Norm Raphael]

> On 01/17/16, Bob Koehler via Info-vax<info-vax at info-vax.com> wrote:
 > 
> In article <c82c8648-aa2e-4f2c-b468-c0e9e1bd9214 at googlegroups.com>, Hans Vlems <hvlems at freenet.de> writes:
> > For a brief period the 8650 was a very fast VAX. With a slow UNIBUS. The VA=
> > X 8550 corrected that. We added a refurbished 8650 to a cluster which had t=
> > wo 11/750's and an 8250 :-). Users complained loudly about missed coffee br=
> > eaks, jobs that ran for 15 minutes plus on the 11/750 completed in 60 secon=
> > ds.
> > Never experienced that kind of performance impact ever since.
> > Hans
> 
 > Never had a VAX 8000. But had a whole bunch of MacroVAXen when we
 > purchased the smallest, slowest Alpha DEC produced.
> 
> Now that's a performance impact.
> 

 On a related note, at one time we had 5 VAX/11-750s (Comets) Each with eight (8, count 'em, 8) 
DZ's.  There were 2 "in" each system, and 6 more mounted on a wall. They were set at 2400 baud.
We replaced them with Able Attach boards - remember Ken (Ethernet will never keep up) Omohundro -  
set at 9600 baud.  User screens that used to paint in 8 seconds began to paint in 2 seconds, and the 
workforce accused management of attempting a speedup as an 8-hour-day's work now took about 
6 hours to complete. Not all improvements are viewed equally by everyone. They adapted.


As an aside, I was once approached to buyacceleration software for the 750's which would have 
increased them 15% from .6 Vups to .69 Vups.  My answer was of course that I wanted to replace 
them, not prolong their use, and we did.  

 
 Norman F. Raphael
"Everything worthwhile eventually 
degenerates into real work." -Murphy