20 Nov
2010
20 Nov
'10
3:49 a.m.
A while back, part of an AMD 2900 development
system showed up on eBay. My copies of the software
surfaced today, so the .imd images are up under
http://bitsavers.org/bits/AMD
They are CP/M 1.4, as I recall.
21 Nov
21 Nov
2:10 a.m.
Interesting stuff.
I have the Mick and Brick book, as well as Donnamaie White Bitslice Design
That one is online:
http://www.donnamaie.com/
I'd be interested in hearing from anybody on their 2900 projects.
Randy
Date: Fri, 19 Nov 2010 19:49:50 -0800
From: aek at bitsavers.org
To: classiccmp at classiccmp.org
Subject: AMSYS29 floppies
A while back, part of an AMD 2900 development
system showed up on eBay. My copies of the software
surfaced today, so the .imd images are up under
http://bitsavers.org/bits/AMD
They are CP/M 1.4, as I recall.
4:20 a.m.
On 11/20/2010 7:10 PM, Randy Dawson wrote:
Interesting stuff.
I have the Mick and Brick book, as well as Donnamaie White Bitslice Design
That one is online:
http://www.donnamaie.com/
I'd be interested in hearing from anybody on their 2900 projects.
Slow! I got the Proms burned ( with a minor bug ), but the rest of the
hardware needs to be fleshed out on paper. I will get back to the
hardware after the new year when I have cash to order the IC's. The
pcbs late spring. I have no rush since I have to write some software
for the machine on the PC so I have a assembler running and maybe a
small compiled language so I can write a simple OS. This is 12/24 bit
CPU with 64Kb of address space that falls in the SMALL computer category.
Randy
Ben.
PS. Any one know of a good brand of serial ( rs232 ) to ISB converter?
4:42 p.m.
On 11/20/10 11:20 PM, ben wrote:
You're burning PROMs but the rest of the computer isn't designed yet??
-Dave
--
Dave McGuire
Port Charlotte, FL
I'd be
interested in hearing from anybody on their 2900 projects.
Slow! I got the Proms burned ( with a minor bug ), but the rest of the
hardware needs to be fleshed out on paper. I will get back to the
hardware after the new year when I have cash to order the IC's. The
pcbs late spring. I have no rush since I have to write some software
for the machine on the PC so I have a assembler running and maybe a
small compiled language so I can write a simple OS. This is 12/24 bit
CPU with 64Kb of address space that falls in the SMALL computer category.
8:38 p.m.
On 11/21/2010 9:42 AM, Dave McGuire wrote:
On 11/20/10 11:20 PM, ben wrote:
You're burning PROMs but the rest of the computer isn't designed yet??
The ALU and Front panel have the schematics and the PCB's done on the
computer.The Memory and I/O card have yet to be finalized. I bought a
new computer and had to wait for a *REAL* keyboard I ordered on line
so I can *hunt and peck* on this thing.
I'd
be interested in hearing from anybody on their 2900 projects.
Slow! I got the Proms burned ( with a minor bug ), but the rest of the
hardware needs to be fleshed out on paper. I will get back to the
hardware after the new year when I have cash to order the IC's. The
pcbs late spring. I have no rush since I have to write some software
for the machine on the PC so I have a assembler running and maybe a
small compiled language so I can write a simple OS. This is 12/24 bit
CPU with 64Kb of address space that falls in the SMALL computer category. -Dave
{click click click snap enter}Ben.
9:16 p.m.
On 11/21/10 3:38 PM, ben wrote:
You're burning PROMs but the rest of the computer isn't designed yet??
The ALU and Front panel have the schematics and the PCB's done on the
computer.The Memory and I/O card have yet to be finalized. I bought a
new computer and had to wait for a *REAL* keyboard I ordered on line
so I can *hunt and peck* on this thing.
Ahh, gotcha. Good luck with your design.
I worked on a commercial (well, "commercial academic", a NASA-funded
project at a university) design that used an Am2910 microsequencer, and
that got me interested in 2900s. Well, that and seeing rows of them on
the FP11-A in my PDP-11/34, and several third-party disk and tape
controllers. ;) I've always wanted to do an Am2900-based design, and I
will if I can ever get some free time.
I do have an Am2901 eval board which is pretty neat.
-Dave
--
Dave McGuire
Port Charlotte, FL
I'd be
interested in hearing from anybody on their 2900 projects.
Slow! I got the Proms burned ( with a minor bug ), but the rest of the
hardware needs to be fleshed out on paper. I will get back to the
hardware after the new year when I have cash to order the IC's. The
pcbs late spring. I have no rush since I have to write some software
for the machine on the PC so I have a assembler running and maybe a
small compiled language so I can write a simple OS. This is 12/24 bit
CPU with 64Kb of address space that falls in the SMALL computer
category. 
22 Nov
22 Nov
2:26 p.m.
From: mcguire at neurotica.com
---snip---
I worked on a commercial (well, "commercial academic", a NASA-funded
project at a university) design that used an Am2910 microsequencer, and
that got me interested in 2900s. Well, that and seeing rows of them on
the FP11-A in my PDP-11/34, and several third-party disk and tape
controllers. ;) I've always wanted to do an Am2900-based design, and I
will if I can ever get some free time.
I do have an Am2901 eval board which is pretty neat.
Hi
I have a S100 board that is said to be a math processor board
that has a single 2901 on it. I have the manual as well.
My Nicolet digital averager scope has several 2901s in it.
They used these just about anywhere one needed a little
more horse power.
Dwight
2:37 p.m.
On 11/22/10 9:26 AM, dwight elvey wrote:
Oh THAT sounds like fun!
I worked on a
commercial (well, "commercial academic", a NASA-funded
project at a university) design that used an Am2910 microsequencer, and
that got me interested in 2900s. Well, that and seeing rows of them on
the FP11-A in my PDP-11/34, and several third-party disk and tape
controllers. ;) I've always wanted to do an Am2900-based design, and I
will if I can ever get some free time.
I do have an Am2901 eval board which is pretty neat.
I have a S100 board that is said to be a math processor board
that has a single 2901 on it. I have the manual as well. My Nicolet digital averager scope has several 2901s
in it.
They used these just about anywhere one needed a little
more horse power.
Yep. Amazingly flexible chips. Even the CPU of the KS10
(DECsystem-2020) is built around them. I think the VAX-11/730 CPU is as
well, if I recall correctly.
-Dave
--
Dave McGuire
Port Charlotte, FL
9:12 p.m.
I have a S100 board that is said to be a math
processor board
that has a single 2901 on it. I have the manual as well.
I find that strange. A single 4 bit slice is not that useful. I asusme
there arten;t 2903s as well :-)
My Nicolet digital averager scope has several 2901s
in it.
They used these just about anywhere one needed a little
more horse power.
Althoguh AFAIK DEC never used them in a PDP11 CPU. There were, of
course, used i nteh floating point processors for some PDP11s, and in the
VAX11/730
Three rivers diddn't use the 2901 in the PERQ, but they did use the 2910
sequencer. I suspect they regretted that decision, it made sense in the
origianl PERQ CPU with its 4K control store (the 2910 has 12 address
outputs), but as the 2910 can't be easily extended ot a larger address
width (unlike the 2909 or 2911 4-bit sequencer slices), all later PERqs
with 16K control stores had a circuit commonly known as the '2 bit
kjludge' (pun totally intentional!)
Apart from the obvious disk and tape controllers in some of my PDP11s,
other machines I've got that use the 2900 series are the Philips P854 (the
P800 CPU in 2900s, basically), a Xerox Daybreak, and the high-speed
language processor in my HP9845. I think there's a 2910 sequencer
(although no 2901s) on the 'transform sequencer bord' in my I2S model 75
image display, but I would have to check
-tony
10:32 p.m.
From: ard at p850ug1.demon.co.uk
Subject: Re: AMSYS29 floppies
To: cctalk at classiccmp.org
Date: Mon, 22 Nov 2010 21:12:41 +0000
I had a 2900 based system for a bit, I was the sales rep for Superset. Wish I had this
machine today!
It was a graphics system, and at the time pretty well advanced. It was a 48 bit
architecture and I remember that it was sort of like a DSP instruction set, in that each
48 bit instruction was op code, two operands and a result. It was designed as a FORTRAN
engine to host a mainframe application, PLOT3D.
We carried it around town, sort of a cube 3'x3' with handles on the side. NASA
guys loved it, but never bought one.
I remember the founder of this company well, he wrote PLOT3D which I understand he later
sold to SAS, as SASGRAPH
I found this article in Dr Dobbs by him:
Ian Hirschon, "Personal Supercomputing: Cray's ideas turn a PC into a
virtual-memory 64-bit supercomputer", Dr. Dobb's Journal, Jun. 1992.*
Last I heard, the box did find a home in the publishing world, with its ability to handle
very large bitmap images.
Anybody else recall this machine?
Randy
I have a S100 board that is said to be a math
processor board
that has a single 2901 on it. I have the manual as well.
I find that strange. A single 4 bit slice is not that useful. I asusme
there arten;t 2903s as well :-)
My Nicolet digital averager scope has several
2901s in it.
They used these just about anywhere one needed a little
more horse power.
Althoguh AFAIK DEC never used them in a PDP11 CPU. There were, of
course, used i nteh floating point processors for some PDP11s, and in the
VAX11/730
Three rivers diddn't use the 2901 in the PERQ, but they did use the 2910
sequencer. I suspect they regretted that decision, it made sense in the
origianl PERQ CPU with its 4K control store (the 2910 has 12 address
outputs), but as the 2910 can't be easily extended ot a larger address
width (unlike the 2909 or 2911 4-bit sequencer slices), all later PERqs
with 16K control stores had a circuit commonly known as the '2 bit
kjludge' (pun totally intentional!)
Apart from the obvious disk and tape controllers in some of my PDP11s,
other machines I've got that use the 2900 series are the Philips P854 (the
P800 CPU in 2900s, basically), a Xerox Daybreak, and the high-speed
language processor in my HP9845. I think there's a 2910 sequencer
(although no 2901s) on the 'transform sequencer bord' in my I2S model 75
image display, but I would have to check
-tony
10:35 p.m.
On 11/22/10 4:12 PM, Tony Duell wrote:
Well, 2901s were fairly expensive chips in those days, and it's
certainly possible to do wider-than-four-bit math in a four-bit CPU. :)
I have a
S100 board that is said to be a math processor board
that has a single 2901 on it. I have the manual as well.
I find that strange. A single 4 bit slice is not that useful. I asusme
there arten;t 2903s as well :-) Althoguh AFAIK DEC never used them in a PDP11 CPU.
There were, of
course, used i nteh floating point processors for some PDP11s, and in the
VAX11/730
And the KS10 as well. I wonder if there are any others; I don't
think so offhand.
One of my favorite uses of Am2901s is in the FPF11 floating-point
processor. For the non-PDP11-savvy, the FPF11 is a strange board that
can be used in both Qbus and Unibus systems by changing jumpers near the
card-edge connectors. It uses a 40-pin ribbon cable to plug into a
microm DIP socket in an F11-based PDP-11 (11/23 or 11/24).
-Dave
--
Dave McGuire
Port Charlotte, FL
23 Nov
23 Nov
6:40 p.m.
On 11/22/10 4:12 PM, Tony Duell wrote:
Well, 2901s were fairly expensive chips in those days, and it's
certainly possible to do wider-than-four-bit math in a four-bit CPU. :)
Indeed it is (heck, there are bit-serial CPUs at the hardware level). But
I am still suprised. Is a 2901 doing floating-point maths really faster
than an 8-bit microprocessor (I am thinking of, say, a PCB with a
dedicated Z80 or similar on it).
Using a 2901 as a nubble-serial processor seems strange because the 16
internal registers are not going to be that useful (64 bits of storage,
OK, you could hold 2 32 bit numbers, I suppose). I would assume there had
to be some external register RAM, in which cae, why use thge 2901? Why
not just use an ALU chip. like the 74S181?
Does the manual for htis board contain schematics? Microcode sources?
Bow much control store is there, is it ROM or RAM, and what is the sequencer?
-tony
I have
a S100 board that is said to be a math processor board
that has a single 2901 on it. I have the manual as well.
I find that strange. A single 4 bit slice is not that useful. I asusme
there arten;t 2903s as well :-)
9:01 p.m.
Hi Tony
I'll have to dig it out. It is in storeage right now
and I'll not get to it for at least 2 weekends.
Dwight
From: ard at p850ug1.demon.co.uk
Subject: Re: AMSYS29 floppies
To: cctalk at classiccmp.org
Date: Tue, 23 Nov 2010 18:40:05 +0000
On 11/22/10 4:12 PM, Tony Duell wrote:
Well, 2901s were fairly expensive chips in those days, and it's
certainly possible to do wider-than-four-bit math in a four-bit CPU. :)
Indeed it is (heck, there are bit-serial CPUs at the hardware level). But
I am still suprised. Is a 2901 doing floating-point maths really faster
than an 8-bit microprocessor (I am thinking of, say, a PCB with a
dedicated Z80 or similar on it).
Using a 2901 as a nubble-serial processor seems strange because the 16
internal registers are not going to be that useful (64 bits of storage,
OK, you could hold 2 32 bit numbers, I suppose). I would assume there had
to be some external register RAM, in which cae, why use thge 2901? Why
not just use an ALU chip. like the 74S181?
Does the manual for htis board contain schematics? Microcode sources?
Bow much control store is there, is it ROM or RAM, and what is the sequencer?
-tony I have a
S100 board that is said to be a math processor board
that has a single 2901 on it. I have the manual as well.
I find that strange. A single 4 bit slice is not that useful. I asusme
there arten;t 2903s as well :-)
4927
days inactive
4930
days old
12 comments
6 participants
participants (6)
-
aek＠bitsavers.org -
ard＠p850ug1.demon.co.uk -
bfranchuk＠jetnet.ab.ca -
dkelvey＠hotmail.com -
mcguire＠neurotica.com -
rdawson16＠hotmail.com