> -----Original Message-----
> From: cctalk [mailto:cctalk-boun...@classiccmp.org] On Behalf Of Charles
> Anthony via cctalk
> Sent: 27 July 2017 21:22
> To: Ed Sharpe <couryho...@aol.com>; General Discussion: On-Topic and
> Off-Topic Posts <cctalk@classiccmp.org>
> Cc: Noel Chiappa <j...@mercury.lcs.mit.edu>
> Subject: Re: Honeywall mainframe CPU front panel ID?
> 
> On Thu, Jul 27, 2017 at 11:55 AM, Ed via cctalk <cctalk@classiccmp.org>
> wrote:
> 
> > In looking at our  collection of  panels  in front of us,  the  large group
> > of smaller and   large black panels  seem  to be the same between all op
> > sys  however.   On the   main panel - yes the difference is the APU
> > lights.
> > As  I  work  though the volume of  stuff in warehousing will have to
> > keep an  eye out  for one  with  APU.
> >
> >
> If you know the provenance of a panel (which site it came from), it may be
> possible to work out what model of machine it came from; Multicians.org has
> a list of sites and some information about the sites hardware.
> 
> 
> 
> > Would  all the  processors  have  APU on the panel  or  just  some of
> > them in a cluster ( if that is the correct  term)
> >
> >
> For a multiple CPU Multics, all of the CPUs would have the APU.
> 
> I don't know if CGOS supported multiple CPUs; a quick google didn't turn up
> any examples, but I don't think that there is any hardware reason why they
> couldn't.
> 

GCOS (and the original GECOS) supported multiple CPU's, up to 4 CPUs. I 
remember being told that Littlewoods (who had 2 x 4 CPU systems) noticed that 
one complex was running slowly, and eventually realized it had lost a CPU 
because a switch on one of the panels we have been discussing was set wrongly.


As for the DPS100/200/300 I think this was merely a "re-badging" of the L66 
systems. When I was at NERC our L66/60 single CPU box was upgraded to a 2 x CPU 
DSP300..

> I don't think there is any difference in GCOS/Multics SCUs; later SCU models
> had more memory (Multics: Many Unnecessarily Large Tables In Core
> Simultaneously).
> 
> Some of the IOMs had a "Multics" mode that enabled them to access the
> Multics page tables; it is not clear(to me) that that model was required for
> Multics; I am researching.
> 
> -- Charles


Dave Wade
G4UGM & EA7KAE



Reply via email to