Original 11/74 front panel

[Don North]

Johnny Billquist wrote:
> Don North <ak6dn at mindspring.com> wrote:
>> Found when I was going thru a box of DEC memorabilia ... an original 
>> 11/74 front panel (plexiglass only). You can tell it is from a 'true' 
>> 11/74 as it has front panel positions for the commercial instruction 
>> set microengine (CIS uADRS, CIS DECIMAL datapath, and CIS DESCRIPTOR 
>> datapath).
>>
>> Here's a pix:   http://www.ak6dn.com/stuff/1174.jpg   (not the best 
>> quality, done on a scanner, not by camera). I need to figure out a 
>> way to get the LED positions to light up. Maybe I'll build a LED 
>> board and hook it up to my 11/34...
>
> Very nice!
>
> However, a few things I've wondered about for a long time.
> This is obviously a panel for a corporate cabinet. I've also seen a 
> picture of an alleged 11/74 that was in a corporate cabinet.
>
> However, the one picture I've seen in real life (as opposed to on the 
> internet) of an 11/74, it was in full height 19" cabinets with the 
> traditional maroon/red color scheme.
The 11/74 was basically an 11/70 as the core CPU (a couple of CPU boards 
were actually updated to provide hooks for the CISP) and a revised 
backplane to handle the multi-card CIS 11/74 option. The 11/74 used the 
same physical chassis as the 11/70, so it could easily be mounted 
anywhere a classic 11/70 could be.
>
> Also, I am under the impression that a CIS was never completed. That 
> would have required the KB11-E. The only 11/74 CPU I know was done was 
> the KB11-Cm, which is a modified KB11-C (aka 11/70). This won't allow 
> you to add a CIS option. No place nor wiring for it.
The multiprocessor version of the 11/70 was not *really* the 11/74. The 
11/70MP modified the microcode for the ASRB instruction to define it as 
the atomic lock primitive (forcing a DATIP/DATOB cycle instead of  a 
plain old DATI/DATOB). It also added the IIST (interprocessor interrupt 
and sanity timer) for interprocessor interrupt capability and a watchdog 
timer.

The 11/74 and CISP were most certainly completed. I know first hand 
because I worked on them and was one of the three microcoders that wrote 
about 4K of 96b microcode to implement the PDP-11 commercial instruction 
set. The complete irony is the same day that engineering signed off the 
system to manufacturing, releasing all the hardware and PROM code to 
production, DEC product management killed it. They finally decided that 
high-end PDP-11s should be EOLed in favor of the newly released 
VAX-11/780 (this was in 1979 IIRC). The 11/44 CIS project started up 
shortly after the 11/74 CIS, but it had a much lower performance target. 
It actually did ship to customers.

Part of the problem was that we did too good a job (IMHO anyway :-). 
PDP-11 Cobol performance (the Cobol compiler actually used CIS 
instructions) on an 11/74 was a factor two or better than that of the 
11/780 running the same Cobol program. Marketing could not figure out 
how to push the new 11/780 into commercial accounts when the 11/74 was a 
clear performance winner. Of course the real benefit of the VAX was the 
32b architecture, allowing programs larger than 128KB (I/D) without 
having to resort to overlays (disk and/or memory swapping).
>
> Does anyone know *for sure* that the KB11-E was made, or might this 
> front have been a prototype for a possible KB11-E, but in reality used 
> just with a KB11-Cm?
Yes, the 11/74 KB11-E was finished by engineering (started in the Mill, 
moved to Tewksbury) but abandoned by DEC marketing. Some number of proto 
systems were around for a while being used internally within DEC but I 
don't believe the 11/74 was ever offered externally.

Don North
DEC 1975-1982
>
> (Oh, and for those of you who care... login to mim.update.uu.se as 
> guest/guest, and run RMD...)
>
>     Johnny
>