[NA] Over engineered "improvements"

[flyingron]

I never did much punch-card stuff, other than 2540s, 2501s, the odd other thing here and there. But I do have a 1600 bit 4-wire core memory card from a 2821 hanging on my wall. Suspect it still has a print train image or something stored on it

I did a little in high school. Part of two of my jobs after college was to GET RID of all the card processing equipment in my facility (first the US Army Ballistic Research Laboratory) and then later Rutgers University's computer center.

There used to be a poster kicking around from the computer museum that showed all the various memory technology: vacuum tubes, mercury delay lines, electrostatic memory, core, and semiconductors. A little research showed that all of them had been used at our labs on computers over the years (admittedly some before my time), we had the ENIAC, ORDVAC, BRLESC, and the Denelcor HEP (all purpose built for BRL). We got the last Cyber 7600 ever built. We also had a couple of Crays, countless PDP-11 and VAXes, some BBN C-70s, and Gould SEL machines before the lab started in on SGI and Sun Workstations (about the end of my tenure there in 1987). This was Reagan-era stuff. In a bad year I bought $2 million dollars in computer equipment each year. The last year I was there, I put my name to the $25 million procurement of the Cray2.

 

[DaleB]

I did a little in high school. Part of two of my jobs after college was to GET RID of all the card processing equipment in my facility (first the US Army Ballistic Research Laboratory) and then later Rutgers University's computer center.

There used to be a poster kicking around from the computer museum that showed all the various memory technology: vacuum tubes, mercury delay lines, electrostatic memory, core, and semiconductors. A little research showed that all of them had been used at our labs on computers over the years (admittedly some before my time), we had the ENIAC, ORDVAC, BRLESC, and the Denelcor HEP (all purpose built for BRL). We got the last Cyber 7600 ever built. We also had a couple of Crays, countless PDP-11 and VAXes, some BBN C-70s, and Gould SEL machines before the lab started in on SGI and Sun Workstations (about the end of my tenure there in 1987). This was Reagan-era stuff. In a bad year I bought $2 million dollars in computer equipment each year. The last year I was there, I put my name to the $25 million procurement of the Cray2.

I've probably written this before, but I started out in 1981/82 repairing IBM 360 Model 30 and Model 40 installations. The Army had them mounted in semitrailers (one for the CPU, console, reader/punch and printer and another for 2420 tape and 2314 disk). The Army was one of the very few users, probably the only one in the US, of the S/360 at the time; it had been obsolete for years. When I got out I was working on - well, bloody near any "big iron" IBM made between the mid 60s and the current generation or just behind it. The last stuff I actually laid hands and tools upon were 4380 and 3090 series systems, along with Xerox 9700 and Siemens cold fusion laser printers. After that I shifted into other lines of work, as it was obvious that field engineers were a rapidly vanishing (and much lower paid) breed.

Good times.

 

[flyingron]

I'm not an IBM guy, but I've given a hand on the 43xx line and the 3090. I worked on AIX for the 370 and other processors under contract from IBM (I still have my IBM badge kicking around in a drawer somewhere). I was also on a team that moved a couple of 3705's with their inane use of bus/tag connectors for RS-232 (what a mess). As a university administrator I was responsible for a Hitatchi AS/9000 (we referred to as the Assinine 9000). Did some work with the early TCP/IP for that platform and was responsible for the COMTEN front end and a bunch of 3270 stuff.

Still I never really considered myself an IBM guy. I spent most of the first half of my career on UNIX (primarily DEC) minicomputers and the latter half on PCs. Wrote one of the first internet routers as well.

 

[Rene]

I did a little in high school. Part of two of my jobs after college was to GET RID of all the card processing equipment in my facility (first the US Army Ballistic Research Laboratory) and then later Rutgers University's computer center.

There used to be a poster kicking around from the computer museum that showed all the various memory technology: vacuum tubes, mercury delay lines, electrostatic memory, core, and semiconductors. A little research showed that all of them had been used at our labs on computers over the years (admittedly some before my time), we had the ENIAC, ORDVAC, BRLESC, and the Denelcor HEP (all purpose built for BRL). We got the last Cyber 7600 ever built. We also had a couple of Crays, countless PDP-11 and VAXes, some BBN C-70s, and Gould SEL machines before the lab started in on SGI and Sun Workstations (about the end of my tenure there in 1987). This was Reagan-era stuff. In a bad year I bought $2 million dollars in computer equipment each year. The last year I was there, I put my name to the $25 million procurement of the Cray2.

Seymour Cray 1925 - 1988

Built a punch tape Morse Code recorder, reader, sender, receiver in his parents' Chippewa Falls, WI basement (using a crystal radio and his erector set), in 1935, at the age of 10 . . .. https://cray-history.net/2021/07/28/about-the-chippewa-falls-museum-of-industry-and-technology/

 

[Bill]

I worked for Big blue for a while, I was an engineer/designer and designed the internal layouts/circuits for the multi layer alumina and glass ceramic substrates used in the IBM 9121 style TCM modules. See “ceramic substrate” below.

 

[flyingron]

Seymour Cray 1925 - 1988
View attachment 138389
Built a punch tape Morse Code recorder, reader, sender, receiver in his parents' Chippewa Falls, WI basement (using a crystal radio and his erector set), in 1935, at the age of 10 . . .. https://cray-history.net/2021/07/28/about-the-chippewa-falls-museum-of-industry-and-technology/

“Seymour said he thought it was odd that Apple bought a Cray to design Macs because he was using Macs to design Crays. He sent me his designs for the Cray 3 in MacDraw on a floppy."

Oddly, the XMP that was supposed to go to Apple got usurped by BRL when the Army decided they were behind the AF and Navy in the supercomputer game and needed the next one off the line. Here's a much younger me standing inside of it...

 

[Bill]

Seymour Cray 1925 - 1988

Ha, streams cross again. Steve Chen worked for Cray and was the lead designer on the X and Y MP supercomputers. Chen left Cray, and formed a partnership with IBM on a company called Supercomputer Systems Inc based in Eau Claire, WI.

I had just rotated off a major project at IBM, and was assigned to TDY in Eau Claire, and I spent the better part of a year working with their engineers on multi-layer ceramic substrate design.

Unfortunately, the project ran out of money before it got to market, and the project was dissolved, Still, interesting times working with those guys.

 

[flyingron]

The BRL X-MP occupied the same spot that my first supercomputer project had... the Denelcor HEP. The Army bought it and then couldn't figure out what they wanted to do for software. My boss and mentor Mike Muuss gave his standard, "We could put UNIX on it," and they didn't have a better idea, so they let us do it. Also after we got it running, we found that the I/O startup (which used an aptly named feature called "The Low-Speed Bus") was very limited (10 i/os a second). Me and HEP designer Burton Smith redesigned the control system literally on napkins at the local Golden Corral building it out of spare parts (an additional memory switch node) and a PDP-11 I had lying around. I wrote the software changes for both the HEP and the PDP-11 and we got supercomputer class I/O going to match the processing power.

 

[Rene]

The BRL X-MP occupied the same spot that my first supercomputer project had... the Denelcor HEP. The Army bought it and then couldn't figure out what they wanted to do for software. My boss and mentor Mike Muuss gave his standard, "We could put UNIX on it," and they didn't have a better idea, so they let us do it. Also after we got it running, we found that the I/O startup (which used an aptly named feature called "The Low-Speed Bus") was very limited (10 i/os a second). Me and HEP designer Burton Smith redesigned the control system literally on napkins at the local Golden Corral building it out of spare parts (an additional memory switch node) and a PDP-11 I had lying around. I wrote the software changes for both the HEP and the PDP-11 and we got supercomputer class I/O going to match the processing power.

 

[DaleB]

I worked for Big blue for a while, I was an engineer/designer and designed the internal layouts/circuits for the multi layer alumina and glass ceramic substrates used in the IBM 9121 style TCM modules. See “ceramic substrate” below.

View attachment 138391

I’ve got an air-cooled TCM from a 4381 around somewhere that now sees use as a paperweight, as well as an occasional reminder of past glory days.

Never did do any software or operational stuff on IBM systems. I was strictly a tool-case-and-scope guy. 360, 370, 303x, 4300, 308x, 3090, s/3 through S/38. The AS400 killed those off right quick, of course.