28 Oct
2014
28 Oct
'14
11:48 p.m.
On 2014-10-28 19:22, Clem Cole wrote:
On Tue, Oct 28, 2014 at 1:46 PM, Johnny Billquist <bqt(a)update.uu.se
<mailto:bqt@update.uu.se>> wrote:
DEC actually made two PDP-11s that were micro programmable. The
11/60 and the 11/03 (if I remember right). DEC never had
microprogramming for the 11/40, but obviously CMU did that.
C.mmp was 11/40E's and C.m* was LSI/11's and both needed them for the
capabilities support. I never really got to mess with the WCS units -
although we learned about them (along with ISPL/ISPS in courses), I did
hack on the OS and in user space of both systems - which was a wonderful
experience. It was how I learned about capabilities which I still have
soft spot. But around that time, I was also introduced to this strange
new system language and system and started to get paid better as a
programmer for a group using it. I never went back ;-)
Yeah, it was the C.mmp I was thinking of when I said that CMU obviously
had to have been playing at this.
BTW: the 780 & 750 had ustore but it was not user
documented and the
tools were internal. Paul Guilbo wrote much of both and later would
write the uCode for the Masscomp FPU and APU. Paul was bitching about
the great tool(s) they had had at DEC, so one weekend two of us on the
SW team got sick of his bitching a couple of us hacked up a uCode
assembler in the same key in Yacc/lex/C (not BLISS ;-).
:-)
As someone else mentioned, the 11/780 had a separate product for user
written microcode. I think it actually also included a different board
for the CPU, with more memory for the microcode. So it must have been
documented externally somehow, somewhere.
The 11/750 was not documented for external use, I think. However, I have
some vague memory of seeing something about user written microcode for
it as well.
And of course, Ultrix had a microcode patch for the 11/750, which fixed
some bugs in a couple of instructions. This microcode patch is still
included in the NetBSD/vax distribution. :-)
The 86x0 machines always loaded microcode from the FE RL02, and there is
documentation on that microcode available today, although it was DEC
internal at the time. And I still happen to have access to an 8650. But
no time to actually try and play with the microcode. And that machine is
rather complex as well. That's when things started to get pipelined and
other stuff that makes things much more difficult to fool around with.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt(a)softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol