From Barton.Bruce@camb.com Tue Aug 23 08:04:57 PDT 1994
Article: 2817 of vmsnet.pdp-11
Path: nntp-server.caltech.edu!elroy.jpl.nasa.gov!swrinde!howland.reston.ans.net!europa.eng.gtefsd.com!MathWorks.Com!noc.near.net!eisner!bruce
From: Barton.Bruce@camb.com (Barton F. Bruce / CCA)
Newsgroups: vmsnet.pdp-11,alt.sys.pdp11
Subject: Re: PDP-11/73 Cabling Questions
Message-ID: <1994Aug22.212256.5026@eisner>
Date: 22 Aug 94 21:22:56 -0400
References: <32peitINNp95@umbc8.umbc.edu> <CuMLBI.HBy@msage.com> <32qtl9INN2sm@umbc8.umbc.edu> <1994Aug17.002524.1@spcvxb.spc.edu> <CuoMoK.Kyx@world.std.com>
Organization: Cambridge Computer Associates, Inc.
Lines: 26
Xref: nntp-server.caltech.edu vmsnet.pdp-11:2817 alt.sys.pdp11:296

In article <CuoMoK.Kyx@world.std.com>, mbg@world.std.com (Megan) writes:
> 
> Another point about the 11/83 - it is additionally defined as a
> KDJ11-B with PMI memory.  Even if you have a KDJ11-B (18mhz and FPA),
> if the system doesn't have PMI memory, it won't be considered an
> 11/83 by system software.
> 
> This is what we in RT-11 Engineering were told by the hardware group
> that developed the board.  That's why if you look in the sources
> for RESORC, you will find code which flushes the cache hit/miss
> register, then does a couple of NOPs, which on a system with PMI
> will cause some ones to be shifted in.  If the hit/miss register
> is non-zero after the NOPs, it's an 11/83 - with PMI.

I thought ALL of the quad wide 11/73 class boards were PMI capable
and that DEC simply did not want you to know that by moving the CPU
down a few slots and stuffing the PMI memory in ahead of it (all
on a Q-CD backplane, of course), you got the PMI functionality.

Otherwise sticking vanilla 22 bit Q bus memory after the processor gave you 
a 'non-PMI' machine.

There are some Tech OEM "Micronotes" on Decuserve that Ozzi Perez kindly 
provided on a TK50 ages ago that cover some of this.

There are lots of 73s out there not in the first slot.


From mbg@world.std.com Tue Aug 23 08:05:06 PDT 1994
Article: 2818 of vmsnet.pdp-11
Newsgroups: vmsnet.pdp-11,alt.sys.pdp11
Path: nntp-server.caltech.edu!elroy.jpl.nasa.gov!swrinde!cs.utexas.edu!uunet!world!mbg
From: mbg@world.std.com (Megan)
Subject: Re: PDP-11/73 Cabling Questions
Message-ID: <Cuyyrv.Mu4@world.std.com>
Organization: The World Public Access UNIX, Brookline, MA
References: <32peitINNp95@umbc8.umbc.edu> <CuMLBI.HBy@msage.com> <32qtl9INN2sm@umbc8.umbc.edu> <1994Aug17.002524.1@spcvxb.spc.edu> <CuoMoK.Kyx@world.std.com> <1994Aug22.212256.5026@eisner>
Date: Tue, 23 Aug 1994 03:52:42 GMT
Lines: 20
Xref: nntp-server.caltech.edu vmsnet.pdp-11:2818 alt.sys.pdp11:297

Barton.Bruce@camb.com (Barton F. Bruce / CCA) writes:

>I thought ALL of the quad wide 11/73 class boards were PMI capable
>and that DEC simply did not want you to know that by moving the CPU
>down a few slots and stuffing the PMI memory in ahead of it (all
>on a Q-CD backplane, of course), you got the PMI functionality.

	That's right.  They ALL are PMI capable.  I've converted
	many machines (including some at DECUS, back when I was
	going) from 11/73B to 11/83 simply by CORRECTLY ordering
	the boards.

>Otherwise sticking vanilla 22 bit Q bus memory after the processor gave you 
>a 'non-PMI' machine.

	Exactly ... what RT refers to as an 11/73B

					Megan Gentry
					Former RT-11 Developer



