Amsterdam Compiler Kit for Cray X-MP | Hacker News Reader

Amsterdam Compiler Kit for Cray X-MP

github.com

93 pointsbasementcat11 comments

30+ years ago I did a week’s “work experience” at the UK offices of Cray Research in Bracknell (and, shockingly, nobody had ever approached them before to do this).

I spent a couple of days with the software support team and was given an account on a UNICOS-running X-MP (hostname was either “forest” or “wind” - I specifically remember the second because the motd said “if you have problems with wind, please contact [redacted]” and that made my 16yo brain chuckle). Anyway, my benchmarking program was to calculate all the factorials up to 100!, and then repeat the process a lot. Fibbonacci, as given in the README, seems like more fun ;)

Hopefully you seldom broke wind while you were there.

Note that this generates executables for COS, the batch processing operating system and not UNICOS, the System V UNIX port.

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

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

This is awesome! I recovered the only copy of COS (http://www.chrisfenton.com/cos-recovery/), but we never really had a way to use it.

One super upvote!I used used this cray ack to some demo code. I really admire kej715 for adding a backend to ack just from reading specification and limited access to hardware as there are no loving cray 1's anymore. I had to try it! Did't build at first so I even contributed a tiny fix to make it work on my sys.

This is also the guy who wrote a DtCyber that emulates many vintage super computers and stuff. Pretty cool! Thank you man!

Can someone please explain what this is good for and whether it is purely of historical interest? It sounds interesting but the readme isn't shedding much light on it.

ACK is a compiler suite (perhaps analogous to gcc). Kevin Jordan maintains a fork of ACK that generates code for the Cray X-MP, a member of a family of famous vector supercomputers. As much of the original software for the Cray family has been lost, this ACK fork may be the only remaining way to program a Cray X-MP (running COS) in a high level language.

https://en.m.wikipedia.org/wiki/Amsterdam_Compiler_Kit

https://en.m.wikipedia.org/wiki/Cray_X-MP

You may be able to run generated executables in an emulator or online.

http://www.nostalgiccomputing.org/

https://github.com/andrastantos/cray-sim

Thanks!

> Currently, this fork builds and runs successfully on MacOS and Linux.

Wow, does that mean that my Mac/Linux box has a Cray X-MP hidden inside it somewhere? I just wanted to start whining that I don't have a Cray...

Finally I can put that Cray X-MP in my basement to work.

Oh wow, I could compile Mitch Bradley's very portable CForth to run on the Cray finally! I wanted to do that from the first time I ever saw a Cray.

https://github.com/MitchBradley/cforth

My "Evil Software Hoarder" friend Mike Gallaher and I from UniPress Software visited Apple once so he could compile and install the latest version of UniPress Emacs on their Cray. I think it was their purple Cray XMP/28 running UNICOS. I don't remember which compiler it used, though.

https://wiki.c2.com/?AppleCrayComputer

https://cray-history.net/2021/07/16/apple-computer-and-cray-...

https://i0.wp.com/cray-history.net/wp-content/uploads/2021/0...

https://retrocomputingforum.com/t/apples-purple-cray-hooked-...

UniPress ported and sold a commercial version of the "Extended Amsterdam Compiler Kit" for Andrew Tanenbaum for many CPUs and versions of Unix (like they also ported and sold his Unix version of Emacs for James Gosling), so Emacs might have been compiled with ACK on the Cray, but I don't recall.

During the late 80's and early 90's, UniPress's Enhanced ACK cost $9,995 for a full source license, $995 for an educational source license, with front ends for C, Pascal, BASIC, Modula-2, Occam, and Fortran, and backends for VAX, 68020, NS32000, Sparc, 80368, and others, on many contemporary versions of Unix.

Rehmi Post at UniPress also made a back-end for ACK that compiled C to PostScript for the NeWS window system and PostScript printers, called "c2ps", which cost $2,995 for binaries or $14,995 for sources.

Independently Arthur van Hoff wrote a different C to PostScript compiler called "PdB" at the Turing Institute, not related to c2ps. It was a much simpler, more powerful, more direct compiler written from scratch, and it supported object oriented PostScript programming in NeWS, subclassing PostScript from C or C from PostScript. I can't remember how much Turing sold it for, but I think it was less than c2ps.

https://compilers.iecc.com/comparch/article/92-04-041

http://www.aaronjamesrogers.com/misc/hotmix16/vendors/sgi/ap...

https://donhopkins.com/home/archive/NeWS/NeScheme.txt

https://news.ycombinator.com/item?id=10088193

I got to watch Apple's Cray compile Emacs very quickly through the plexiglass viewing window, but they wouldn't let me sit or nap on it, alas.

I knew a guy who worked at one of the national labs that had its own Cray supercomputer, in a computer room with a big observation window that visitors could admire it through, of course (all Crays required observations windows to show them off).

Just before a tour group came by, he hid inside the Cray, and waited for them to arrive. Then he casually strolled out from the back of the Cray, pulling up the zipper of his jeans, with a sheepish relieved expression on his face, looked up and saw the tour group, acted startled, then scurried away.

Reply on news.ycombinator.com