Not any PERQ that I've seen (and I think I've seen all models). The classic PERQs (1, 1a, 2T1, 2T2, 2T4) used 74S181 ALUs and 256 nybble RAMs for the registers (yes, 256 CPU registers). The sequencer was an AMD 2910, which, of curse, is from the same family as the 2901. But the 2910 is not IMHO a true 'slice' in that it can't be extended to <n> bits. It's a 12 bit microcode sequencer (and all PERQs apart from the 1 had 16K of control store, so they had to extend the sequener address by a circuit kmown to all PERQ fanatics as the '2 bit kludge', pun totally intentional). The PERQ ethernet interface and hard disk cotnrollers also used 2910 sequencers with PROMs as the control store. The PERQ 3a (AGW3300) was a 68020-based machine. The graphics processor used a pair of 29116 data path chips (one to calculate addresses, one to do the data processing) along with a very simple sequencer built from TTL. -tony

And another one I've just remembered : Philips P854. The P851, BTW, was built from Philips's own custom bit-slice chip called SPALU (Scratch Pad and Arithmetic Logic Unit) -tony

That, in a convoluted way, has reminded me of a couple more 29xx-based processors The HP9845 Opt 200 High Speed Language Processor contains 4 2901s and a 2910 The HP98770 colour monitor contains a video processor with, IIRC, 3 2903s on it. And the Xerox Daybreak contains 4 2901-like things (maybe 7C901s) and a custom sequencer chip.

Steven N. Hirsch wrote: ... Google says: http://www.500sound.com/SyncII/sync2intro.htm This says the CPU was called "ABLE", and it was programmed in "Scientific XPL".

J. David Bryan wrote: And furthermore: does anyone know of a good source for the 2904 (the shifter/status register extension chip). The usual sources like unicorn, digikey, and bgmicro only offer, if at all the common ALU/sequencers 2901/2903/2909/10/11, but the 2904 seems to be hard to find. Looks as if I'd have to replace it with a CPLD. Holger

On Fri, 30 Dec 2005 18:20:30 -0700, you wrote: If fading memory serves, the Hughes 5118ME mini used 2901s. The 5118ME was (and still is, to some degree) used in Air Defense systems dating from the 1980s, onward. No documentation to back it up, just a vague memory of the times. Mike Loewen mloewen at cpumagic.scol.pa.us Old Technology http://ripsaw.cac.psu.edu/~mloewen/Oldtech/

On Sat, 7 Jan 2006, Bob Shannon wrote: Possibly. I recall seeing references to an older, 3118 machine. The 5118 is a dual processor, real-time computer, and was the central processor in the JSS air defense system, connected to two HMP-1116 minicomputers which handled radar inputs, console displays, printer output and hard disk storage, and a 3rd 1116 which handles the color status monitor. The whole thing connects to a redundant system through one of the 1116s. The 5118 had a wire wrapped backplane. There's an article that mentions the 3118, 4118 and 5118: http://www.aiaa.org/content.cfm?pageid=406&gTable=mtgpaper&gID=67515 Mike Loewen mloewen at cpumagic.scol.pa.us Old Technology http://ripsaw.cac.psu.edu/~mloewen/Oldtech/

Yes, these are clearly relatives! I worked on a combat air traffic control system, run by a 4118. It was a single CPU 18-bit machine, but with dedicated coprocessors for radar video processing and a vector display system for radar consoles. ----- Original Message ----- From: "Mike Loewen" <mloewen at cpumagic.scol.pa.us> To: "General Discussion: On-Topic and Off-Topic Posts" <cctalk at classiccmp.org> Sent: Saturday, January 07, 2006 3:28 PM Subject: Re: AMD bit-slice machines