Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Prime bitslice processor

85 views
Skip to first unread message

Jim Wilcoxson

unread,
Apr 24, 2020, 10:44:01 AM4/24/20
to
Is this the chip that Prime used to implement its CPUs?

http://www.righto.com/2020/04/inside-am2901-amds-1970s-bit-slice.html

Dennis Boone

unread,
Apr 24, 2020, 12:05:02 PM4/24/20
to
> Is this the chip that Prime used to implement its CPUs?
> http://www.righto.com/2020/04/inside-am2901-amds-1970s-bit-slice.html

The 2900 family was used in at least some of the designs.

There's another common ALU, the 74181, which might be a candidate
for the earliest machines.

De

Bob Allison

unread,
Oct 21, 2020, 10:12:38 PM10/21/20
to
On Friday, April 24, 2020 at 10:44:01 AM UTC-4, Jim Wilcoxson wrote:
> Is this the chip that Prime used to implement its CPUs?
>
> http://www.righto.com/2020/04/inside-am2901-amds-1970s-bit-slice.html
The 2901 bit slices were used in the early ecl machines but once we got to the 6000 series we were
designing our own chips using Motorola MCA2500 ECL Macrocell Arrays on all the boards of the CPU.

Daiyu Hurst

unread,
Jul 10, 2021, 10:55:40 PM7/10/21
to
Were the P300 and P400 ECL machines? I hadn't thought so. I have the P300 microcode manual, and we once had the microcode boards for the P400 at Greater Clark County Schools. I remember looking at them, and I'm pretty sure that's what I recall was used.

On Friday, April 24, 2020 at 12:05:02 PM UTC-4, Dennis Boone wrote:
> > Is this the chip that Prime used to implement its CPUs?
> > http://www.righto.com/2020/04/inside-am2901-amds-1970s-bit-slice.html
> The 2900 family was used in at least some of the designs.
>
> There's another common ALU, the 74181, which might be a candidate
> for the earliest machines.

I have the data sheets for the 54181/74181 ALU in my 1976 (second printing) Texas Instruments TTL Data Book. if they're not already online somewhere, let me know, I'll scan the pages.

But I'm a stroked-out old lady; what do I know, lol.

-Dai

Dennis Boone

unread,
Jul 11, 2021, 9:24:39 PM7/11/21
to
> Were the P300 and P400 ECL machines? I hadn't thought so. I have the
> P300 microcode manual, and we once had the microcode boards for the P400
> at Greater Clark County Schools. I remember looking at them, and I'm
> pretty sure that's what I recall was used.

No, pretty sure the first ECL machine was the 9950.

> I have the data sheets for the 54181/74181 ALU in my 1976 (second
> printing) Texas Instruments TTL Data Book. if they're not already
> online somewhere, let me know, I'll scan the pages.

Check http://bitsavers.org/components/ti/_dataBooks/ - I suspect
it's in there.

De

Daiyu Hurst

unread,
Jul 13, 2021, 1:18:10 PM7/13/21
to
Yup. Minus the faux-leather covers. Al seems not to favor the covers.

Dennis Boone

unread,
Jul 13, 2021, 9:18:11 PM7/13/21
to
> Yup. Minus the faux-leather covers. Al seems not to favor the covers.

Given the volume he scans, if it won't go through the sheet feeder,
I suspect they sometimes get skipped. :)

De
Message has been deleted
Message has been deleted

Bob Allison

unread,
Jan 23, 2023, 2:05:07 AM1/23/23
to
Yes, the 9950 and all 9000s were ECL with the 2900 family of bit-slice chips.
There were 6 boards in the CPU: Execute Units E1 & E2, Instruction I-Unit, S-Unit (Cache), CS (Control Store), and MC (Memory Controller).

The 6000 series was ECL with custom Prime-designed Motorola MCA 2500 ECL VLSI.
There were 6 boards in the CPU: the Execution Unit, Instruction/Cache Unit, Control Store, and Memory Controller with multiple VLSI designs on each of the 4 boards that made up the CPU.

The next ECL CPU (The Lion project) was to be with the MCA 25000 VLSI, but that project was canceled after the takeover attempt and the buyout by JH Whitney and DR Holdings.

The 5000 series was with custom Prime-designed CMOS from LSI Logic. The single processor variations had multiple chips for the complete CPU on one board plus slots for memory modules on the same board. The 5370 dual processor version had two sets of CPU chips and a new Memory Controller all on one board with a second board for all the memory modules. Those projects survived the takeover and the 5370 dual CPU was the last machine designed and shipped from Prime.

Daiyu Hurst

unread,
Jun 27, 2023, 2:41:25 PM6/27/23
to
I really thought I remembered seeing the AM2900 in the microcode boards for the P400.

Dennis Boone

unread,
Jun 27, 2023, 10:01:28 PM6/27/23
to
> I really thought I remembered seeing the AM2900 in the microcode boards
> for the P400.

If you compare pdf page 75 here -

https://sysovl.info/pages/blobs/prime/archhw/MAN1857_microcodingCodingHbk_Sep74.pdf

to the function table for the 74181 here -

https://en.wikipedia.org/wiki/74181

they seem to be the same thing.

De

Dennis Boone

unread,
Jun 27, 2023, 10:46:27 PM6/27/23
to
> > I really thought I remembered seeing the AM2900 in the microcode boards
> > for the P400.

I have a board set from a 150. It's a little after the 400, of course,
but the cpu board has several 74s181 chips on board.

https://yagi.h-net.org/p150_cpu_74s181/

De
0 new messages