30 Jun
2010
30 Jun
'10
8:27 a.m.
"Shoppa, Tim" <tshoppa at wmata.com> wrote:
"Walter F.J. Mueller" <w.f.j.mueller at
gsi.de> wrote:
Yes, I know.
I'll reply to this one last time, and then I'll give up.
Don North reported that he had been a part of the team that had written
the CIS microcode for the 11/74 CPU.
I commented that I thought the 11/74 CPU had only been a paper product.
Don North also pointed out that marketing "stole" the 11/74 moniker for
the 11/70mP system.
Now, throughout this discussion, we need some way of separating what we
are talking about. DEC internal project papers seems like a good start.
There we have the 11/70mP, which is a modified 11/70 with just the
addition of the ASRB cache bypass and memory interlock, as well as the
cache bypass bit in the PDR, and a cache bypass bit and flush control in
the cache control CSR.
The 11/74 is a total redesign of the 11/70 CPU, with the same
modifications as the 11/70mP, but also the addition of the CIS, removal
of one Massbus, and redesign of a whole bunch of CPU boards, including
removing one clock signal not used, and the addition of new clock
signals and control signals required by the CIS.
I'm only talking CPUs here, not systems.
Another way to name them would perhaps be:
KB11-B - Old 11/70 CPU with synch FPP.
KB11-C - New 11/70 CPU with asynch FPP.
KB11-CM - MP modified KB11-C
KB11-E(?) - The new 11/74 CPU with asynch FPP and CIS.
I seem to remember reading somewhere that the 11/74 CPU were to be
called KB11-E, but I also have this nagging feeling that KB11-E might
have been the 11/44, or possibly the 11/60.
Now, as I myself pointed out, RSX regards the 11/70mP as an 11/74, and
that is also what the CPU identification code in RSX calls it.
But if we call this an 11/74, what shall we call the 11/70 with CIS?
So, for the purpose of this thread, I decided to go with Don Norths
naming, and call the 11/70 modified for multiprocessor operations the
11/70mP. If you look at the picture on your site, Tim, you'll also
notice that the text on the front panel actually says something like
"PDP-11/74 MP". (Not sure about the /74, but you definitely see the
"MP"
part. (http://www.trailing-edge.com/~shoppa/1174Xopen.jpg)
Now, compare that to Don Norths picture of the 11/74 front panel:
http://www.ak6dn.com/stuff/1174.jpg.
> Johnny Billquist <bqt at softjar.se>
wrote:
> > I wasn't aware that any prototypes ever were produced and came as
> > far as being functional. I thought it was just paper work that
> > had bee done.
>
> The 11/74 wasn't marketed, as pointed out in this thread, but a
> few systems were build by DEC. A picture of such 11/74 system
> was made available by Tim Shoppa, see
>
> http://www.trailing-edge.com/~shoppa/1174Xopen.jpg
>
> You'll nicely see the four CPUs.
Yes, I know of these systems. However, that is
not an 11/74 on that
picture, but an 11/70mP. There is a difference...
As pointed out, the 11/70mP was marketed as an 11/74, but it's a
different CPU.
The easiest way to see that this is a picture of
an 11/70mP is by
looking at the lower rotary switch, which only have four positions, and
not eight (which the 11/74 have). So no CIS on this machine.
The only 11/74 picture I've seen so far is
the silk screen panel picture
posted a few days ago. Unfortunately I've already forgotten the name
(I'm lousy with names, sorry) of the person who posted it, and who also
worked on the 11/74 CIS microcode.
The machine on that picture is probably CASTOR::
by the way.
The people who work with/maintain CASTOR:: call it a 11/74, FWIW. 
They never used the term "11/70mP" in front
of me for sure. I would occasionally elicit comments about multiprocessing on 73's or
93's but it always came back to "our 11/74 does it THIS WAY" because that
was the working example.
I'm not disagreeing with you, Tim. I'm just trying to point out that we
have two different CPUs here, one of which I thought was never made, but
Don actually claims that it did exist, even if just as one prototype.
The system was called an 11/74 everywhere, but for the purpose of this
discussion, we need to make a distinction between the CPUs.
I'm not saying that "11/70mP" is wrong,
indeed it's used in some of the drawings and memos to describe what was commonly
called the 11/74.
Yes.
CIS was real important to some DECcies circa late
70's for some Cobol requirement but coming from the real-time side none of us ever
cared. We'd just run across machines that had this unneeded option.
Indeed. And the 11/70 don't have it, nor does the 11/74 systems that
ever were used.
CASTOR:: was 4 CPUs, by the way, while PHEANX:: was only 2, if I
remember right.
Johnny
30 Jun
30 Jun
10:23 a.m.
Johnny Billquist wrote:
Another way to name them would perhaps be:
KB11-B - Old 11/70 CPU with synch FPP.
KB11-C - New 11/70 CPU with asynch FPP.
KB11-CM - MP modified KB11-C
KB11-E(?) - The new 11/74 CPU with asynch FPP and CIS.
I seem to remember reading somewhere that the 11/74 CPU were to be
called KB11-E, but I also have this nagging feeling that KB11-E might
have been the 11/44, or possibly the 11/60.
The 11/44 CPU was a KD11-Z. The 11/60 CPU was a KD11-K.
Now, as I myself pointed out, RSX regards the 11/70mP
as an 11/74, and
that is also what the CPU identification code in RSX calls it.
Since the 11/70mp and 11/74 were never official products, there was a
lot of conflation of the designations. Without the optional CIS,
software can't easily distinguish an 11/70mp from an 11/74, so it
probably simply didn't bother to try. Thus whether software reports the
CPU as an 11/70mp or an 11/74 doesn't really prove much of anything.
But if we call this an 11/74, what shall we call the
11/70 with CIS?
Fantasy? There wasn't such a thing, since there wasn't an 11/70 with a
KB11-E CPU that was necessary to accomodate the KE74-A CIS.
Eric
2:56 p.m.
On 6/30/2010 10:23 AM, Eric Smith wrote:
I believe s/w could have examined PSW<8> the 'CIS Instruction Suspended'
bit in the KB11-E w/o the CIS option installed, and it could have set
that bit to a one, whereas in the 11/70 that bit would have been locked
at zero.
Correct. The only backplane that could accept the KE74-A CIS option was
the '11/74' KB11-E backplane.
Johnny Billquist wrote:
Correct. 11/74 was the KB11-E. 11/45 was the KB11-A, 11/55 was the
KB11-D, just to complete the table.
Another way to name them would perhaps be:
KB11-B - Old 11/70 CPU with synch FPP.
KB11-C - New 11/70 CPU with asynch FPP.
KB11-CM - MP modified KB11-C
KB11-E(?) - The new 11/74 CPU with asynch FPP and CIS.
I seem to remember reading somewhere that the 11/74 CPU were to be
called KB11-E, but I also have this nagging feeling that KB11-E might
have been the 11/44, or possibly the 11/60.
The 11/44 CPU was a KD11-Z. The 11/60 CPU was a KD11-K. Now, as I myself pointed out, RSX regards the
11/70mP as an 11/74,
and that is also what the CPU identification code in RSX calls it.
Since the 11/70mp and 11/74 were never official products, there was a
lot of conflation of the designations. Without the optional CIS,
software can't easily distinguish an 11/70mp from an 11/74, so it
probably simply didn't bother to try. Thus whether software reports
the CPU as an 11/70mp or an 11/74 doesn't really prove much of anything. But if we call this an 11/74, what shall we call
the 11/70 with CIS?
Fantasy? There wasn't such a thing, since there wasn't an 11/70 with
a KB11-E CPU that was necessary to accomodate the KE74-A CIS.
Eric
11:46 a.m.
On 06/30/2010 11:27 AM, Johnny Billquist wrote:
"Shoppa, Tim" <tshoppa at wmata.com>
wrote:
I can't add too much to this regarding what parts and what DEC
designators applied
but here are memories of the time frame.
The first multiprocessor 11/70 was built with existing hardware and a few
wire wrap and jumper mods. Memory said there were 4 total, three inside DEC
and one at CMU that they hacked together possibly with DEC help. It
would evolve
to a design project to make that buildable as marketing felt they could
sell it.
However at the same time VAX/11/780 was real and also the various product
groups were feeling the effects of FCCs new class A and B limits for
RFI/EMI.
That and the high end market had been moving to more addressable memory
for bigger datasets and computationally wider data words as the tasks were
getting bigger. At that time the big calculations that were important were
atomic physics and weather models and both were associated with massive
[by that eras measure] datasets. In many respects the same pressures
repeated
themselves in the 32bit to 64bit evolution [Alpha].
It was my understanding that the 11/70 continued as a grandfathered
EMI and the new multiple cpu died due to EMI issues (plethora of cables
and multiple racks) and it was a faster number cruncher than VAX-11/780.
The VAX had higher potential as the new reigning super minicomputer. It
wasn't long after that I'd seen a VAX-11/782, 785 and VAXclusters.
There were several of the PDP11 flavors that would die or morph as a
result of manufacturing and serviceability issues.
Allison
"Walter F.J. Mueller" <w.f.j.mueller
at gsi.de> wrote:
Yes, I know.
I'll reply to this one last time, and then I'll give up.
Don North reported that he had been a part of the team that had
written the CIS microcode for the 11/74 CPU.
I commented that I thought the 11/74 CPU had only been a paper product.
Don North also pointed out that marketing "stole" the 11/74 moniker
for the 11/70mP system.
Now, throughout this discussion, we need some way of separating what
we are talking about. DEC internal project papers seems like a good
start.
There we have the 11/70mP, which is a modified 11/70 with just the
addition of the ASRB cache bypass and memory interlock, as well as the
cache bypass bit in the PDR, and a cache bypass bit and flush control
in the cache control CSR.
The 11/74 is a total redesign of the 11/70 CPU, with the same
modifications as the 11/70mP, but also the addition of the CIS,
removal of one Massbus, and redesign of a whole bunch of CPU boards,
including removing one clock signal not used, and the addition of new
clock signals and control signals required by the CIS.
I'm only talking CPUs here, not systems.
Another way to name them would perhaps be:
KB11-B - Old 11/70 CPU with synch FPP.
KB11-C - New 11/70 CPU with asynch FPP.
KB11-CM - MP modified KB11-C
KB11-E(?) - The new 11/74 CPU with asynch FPP and CIS.
I seem to remember reading somewhere that the 11/74 CPU were to be
called KB11-E, but I also have this nagging feeling that KB11-E might
have been the 11/44, or possibly the 11/60.
Now, as I myself pointed out, RSX regards the 11/70mP as an 11/74, and
that is also what the CPU identification code in RSX calls it.
But if we call this an 11/74, what shall we call the 11/70 with CIS?
So, for the purpose of this thread, I decided to go with Don Norths
naming, and call the 11/70 modified for multiprocessor operations the
11/70mP. If you look at the picture on your site, Tim, you'll also
notice that the text on the front panel actually says something like
"PDP-11/74 MP". (Not sure about the /74, but you definitely see the
"MP" part. (http://www.trailing-edge.com/~shoppa/1174Xopen.jpg)
Now, compare that to Don Norths picture of the 11/74 front panel:
http://www.ak6dn.com/stuff/1174.jpg.
> Johnny Billquist <bqt at softjar.se>
wrote:
> > I wasn't aware that any prototypes ever were produced and came as
> > far as being functional. I thought it was just paper work that
> > had bee done.
>
> The 11/74 wasn't marketed, as pointed out in this thread, but a
> few systems were build by DEC. A picture of such 11/74 system
> was made available by Tim Shoppa, see
>
> http://www.trailing-edge.com/~shoppa/1174Xopen.jpg
>
> You'll nicely see the four CPUs.
Yes, I know of these systems. However, that is
not an 11/74 on that
picture, but an 11/70mP. There is a difference...
As pointed out, the 11/70mP was marketed as an 11/74, but it's a
different CPU.
The easiest way to see that this is a picture of
an 11/70mP is by
looking at the lower rotary switch, which only have four positions,
and not eight (which the 11/74 have). So no CIS on this machine.
The only 11/74 picture I've seen so far is
the silk screen panel
picture posted a few days ago. Unfortunately I've already forgotten
the name (I'm lousy with names, sorry) of the person who posted it,
and who also worked on the 11/74 CIS microcode.
The machine on that picture is probably CASTOR::
by the way.
The people who work with/maintain CASTOR:: call it a 11/74, FWIW. They never used the term "11/70mP" in
front of me for sure. I would
occasionally elicit comments about multiprocessing on 73's or 93's
but it always came back to "our 11/74 does it THIS WAY" because that
was the working example.
I'm not disagreeing with you, Tim. I'm just trying to point out that
we have two different CPUs here, one of which I thought was never
made, but Don actually claims that it did exist, even if just as one
prototype.
The system was called an 11/74 everywhere, but for the purpose of this
discussion, we need to make a distinction between the CPUs.
I'm not saying that "11/70mP" is
wrong, indeed it's used in some of
the drawings and memos to describe what was commonly called the 11/74.
Yes.
CIS was real important to some DECcies circa late
70's for some Cobol
requirement but coming from the real-time side none of us ever cared.
We'd just run across machines that had this unneeded option.
Indeed. And the 11/70 don't have it, nor does the 11/74 systems that
ever were used.
CASTOR:: was 4 CPUs, by the way, while PHEANX:: was only 2, if I
remember right.
Johnny
3:28 p.m.
... and it was a faster number cruncher than
VAX-11/780.
The VAX had higher potential as the new reigning super minicomputer. ?It
wasn't long after that I'd seen a VAX-11/782, 785 and VAXclusters.
I have to think that these later VAX machines and clusters were
certainly well in motion, even if unannounced, by the time this whole
11/70mp / 11/74 project came out. In a business perspective, it makes
a lot of sense why these weird-11s were not marketed.
I know almost nothing technically about these, so I ask - were these
multi-processor machines testbeds for ideas later seen (or not) in
other machines? Were they pet projects (DECBob)?
--
Will
1 Jul
1 Jul
6:51 a.m.
On 06/30/2010 06:28 PM, William Donzelli wrote:
Well you know the VAX11/70 announcement date it's easy to work backward by
18months to two years to get to a design start date. Since the
announcement was
introduced on October 25, 1977 we are looking at late 1975. There is
much overlap
in he time lines.
Also the supposed picture of the 70mp is the same cab style and colors
of the
post 780 era.
... and it was
a faster number cruncher than VAX-11/780.
The VAX had higher potential as the new reigning super minicomputer. It
wasn't long after that I'd seen a VAX-11/782, 785 and VAXclusters.
I have to think that these later VAX machines and clusters were
certainly well in motion, even if unannounced, by the time this whole
11/70mp / 11/74 project came out. In a business perspective, it makes
a lot of sense why these weird-11s were not marketed.
I know almost nothing technically about these, so I
ask - were these
multi-processor machines testbeds for ideas later seen (or not) in
other machines? Were they pet projects (DECBob)?
VAX SMP machines and later and Clusters of DEC36bit machines may have been.
Allison
--
Will
7:54 a.m.
I'd strongly suggest reading
*Computer Engineering - C. Gordon Bell, J. Craig Mudge, John E. McNamara*
On line at:
http://research.microsoft.com/en-us/um/people/gbell/Computer_Engineering/
As it covers a lot of the history leading up to the 11/70 and VAX as
well as other machines
from 1957 to about 1977. it's straight from the
horses moth info and
also provides the
time lines and a good look at why and what.
It's important to point out that PDP11 was aimed at low to mid size
applications of the time
and VAX was for the bigger jobs and demanding apps. it was the successor
to the PDP-8
for a large segment of computer tasks and it was to be cheaper and
perform better. There
was overlap in that, a great deal of it both in performance and
timeframe. The 11/70 was
for practical cases concurrent with with the VAX11/780 and the 11/74 was
after that. When
you consider the early VAX system also offered compatibility with PDP11
it didn't take a genius
to see that there was customer base market fragmentation over PDP11 or
VAX. The
problem for DEC and the customers was that there were two paths one
being upgrade with
performance and lifespan ahead and the other cheaper but near the end
for how much more
performance it could yield. For many parts of DEC this was a conundrum
as PDP11 was popular
and well established by 1975 but it was clear by 1980 it's ranks would
be relegated to lower
cost medium performance due to technology growth where a machine
embodying more
advanced concepts would have extended growth path. This was not new to
the industry
but DEC was like many in the middle of it trying to figure out the next
direction and those
there were at most starting a third decade since PDP-1 and knew that
change was the only
constant.
It's important to recognize that computer systems in 1978 were more
costly than a car for small
ones and for the midsize PDP11 more than a house, and VAX was in the
price range of a entire
block of houses. In 1978 that dynamic range was 4000 to 1,000,000 plus
dollars. That and the
rapid advancement of computer performance over short periods made it
clear if you weren't
looking ahead you were headed for the wall. Even the PC market
exhibited this and its time
scale was considerably compressed in comparison.
My experiences span 1969 with my first contact with an pdp-8i (BOCES
LIRICS Tymeshare)
with the works through PDP10, PDP11, 8008, 8080, Z80 to current ARM9.
Thats a lot of change.
Allison
On 07/01/2010 09:51 AM, allison wrote:
On 06/30/2010 06:28 PM, William Donzelli wrote:
Well you know the VAX11/70 announcement date it's easy to work
backward by
18months to two years to get to a design start date. Since the
announcement was
introduced on October 25, 1977 we are looking at late 1975. There is
much overlap
in he time lines.
Also the supposed picture of the 70mp is the same cab style and colors
of the
post 780 era.
... and
it was a faster number cruncher than VAX-11/780.
The VAX had higher potential as the new reigning super
minicomputer. It
wasn't long after that I'd seen a VAX-11/782, 785 and VAXclusters.
I have
to think that these later VAX machines and clusters were
certainly well in motion, even if unannounced, by the time this whole
11/70mp / 11/74 project came out. In a business perspective, it makes
a lot of sense why these weird-11s were not marketed.
I know almost nothing technically about these, so
I ask - were these
multi-processor machines testbeds for ideas later seen (or not) in
other machines? Were they pet projects (DECBob)?
VAX SMP machines and later and Clusters of DEC36bit machines may have
been.
Allison
--
Will
30 Jun
30 Jun
6:17 p.m.
allison wrote:
It was my understanding that the 11/70 continued as a
grandfathered
EMI and the new multiple cpu died due to EMI issues (plethora of cables
and multiple racks) and [...]
That sounds *much* more plausible than the problem being that the flat
cables were a maintenance/serviceability issue, as has often been reported.
Of course, with enough effort, they could have solved the problem, but
likely didn't want to for theother reasons you mentioned.
Eric
6:43 p.m.
On 06/30/2010 09:17 PM, Eric Smith wrote:
allison wrote:
They were but, changing it meant making the grandfathering a problem.
There were far to many cables that
could easily be swapped and cause a failure or more commonly a no run
case till it was sorted out.
Around that time a New engineering group CSSE wich was part of Central
engineering but funded by
filed service was in existence to solve manufacturing quality and field
maintainability issues. That was
spawned by FA&T (factory assemble and test) because supposedly
interchangeable parts weren't
so and systems would not work in the field (early 70s). So the cable
issue was very real but it was
not the sole reason.
It was my understanding that the 11/70 continued
as a grandfathered
EMI and the new multiple cpu died due to EMI issues (plethora of cables
and multiple racks) and [...]
That sounds *much* more plausible than the problem being that the flat
cables were a maintenance/serviceability issue, as has often been
reported.
Of course, with enough effort, they could have solved
the problem, but
likely didn't want to for theother reasons you mentioned.
Therein lies the problem, FCC rules of the time meant if you redesign it
your committed
to "fix" the RFI issue. That meant only ECOs and retrofitable options.
The VAX and later
PDP-11 machines shows that packaging impact.
Also for those that were in DEC there was always a bit of a split in
product groups where some
wanted to keep tweaking the design for cost and speed and those that
wanted a whole new
challenge and direction. Some of that shows in the various ways the
PDP-8 and PDP11 were
implemented over the years.
Just trying to keep up with that was a experience. My years in DEC
engineering were among the
best in my career.
Allison
Eric
5054
days inactive
5055
days old
8 comments
5 participants
participants (5)
-
ajp166＠verizon.net -
ak6dn＠mindspring.com -
bqt＠softjar.se -
eric＠brouhaha.com -
wdonzelli＠gmail.com