Article View: alt.folklore.computers
Article #316CDC 6600 infection
From: fu... Mark Fulk
Date: Thu, 14 Dec 1989 06:24
Date: Thu, 14 Dec 1989 06:24
43 lines
2899 bytes
2899 bytes
I learned this one while working at the Courant Institute of Mathematical Sciences, New York University: Courant had vast sums of money from the Atomic Energy Commission (later ERDA, then DOE). They ran a major computer installation which featured one of the first CDC 6600's; in fact, serial number 3. Many wonderful things were done on this machine, and many to it. Ever hear of the supercritical wing? Well, the CDC had a bad habit of crashing at 3:00 or so weekdays. It was fairly regular about this; every few weeks, at 3:00, it would crash. The repairman would come in, find one of the "cordwood" PC cards burnt out, replace it, and things would be back to normal for a while. But everyone worried about the mystery of the regular crash. Turns out that, because the machine was number 3, the backplane had been hand wire-wrapped. The technicians had left fingerprints all over the place. A fungus was growing on the backplane, apparently fed by the oils in the fingerprints. Most of the time, the air in the room was very dry, and the fungus didn't grow. But when the operators came in on weekdays, the moisture in their breath was enough for the fungus to grow a bit. Eventually it would short a couple of pins (remember, the 6600's circuitry was run at the ragged edge) causing something to burn out. Of course, that bit of fungus would get incinerated too. An application of spray fungicide cured the problem. --------------------- This machine is the subject of a bazillion stories, as are some of the people around it. It's the machine that Ken Kesey and the merry pranksters threatened to blow up (hence, despite programming on it for a year, I never saw it; the floor it was on was locked up). Jacob Schwartz, my old boss, was famous for inventing the first way to save one register when you couldn't touch the rest: a long series of test and branch instructions (the Courant 6600 didn't have an exchange jump instruction; that came later). Garabedian and Jameson did the first numerical simulations of the supercritical wing on it; that idea later resulted in 30% fuel savings for jet engines. Charles Peskin used it to simulate blood flow in a dog's heart. I used UT LISP on it; the only LISP with CONS, CAR, CDR, and CSR (cooser); addresses were 18 bits, words were 60 bits, so they put another pointer in a CONS cell. Mark
Technical Details
Message-ID:
<1989Dec14.062403.16907@cs.rochester.edu>Path:
mbox-mo1.pugleaf.net!gmdzi!unido!mcsun!uunet!wuarchive!udel!rochester!fulkShow Full Headers
Complete Article Headers
Path: mbox-mo1.pugleaf.net!gmdzi!unido!mcsun!uunet!wuarchive!udel!rochester!fulk From: fu...@cs.rochester.edu (Mark Fulk) Newsgroups: alt.folklore.computers Subject: CDC 6600 infection Message-ID: <1989Dec14.062403.16907@cs.rochester.edu> Date: 14 Dec 89 06:24:03 GMT Reply-To: fu...@cs.rochester.edu (Mark Fulk) Organization: U of Rochester, CS Dept, Rochester, NY Lines: 43 Posted: Thu Dec 14 07:24:03 1989 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit