CDC 6500 supercomputer at the Living Computers Museum + Labs

[@lawrencedoliveiro9104]

6:34 Dropping your cards wasn’t such a big calamity, if you were well-organized. This is why languages of the era, like FORTRAN and COBOL, would ignore anything punched in columns 72-80; these were reserved for serial numbers for ordering the cards. There were special card-sorting machines, so you just had to drop your deck into them, specify the sort columns, and they would come out with the right order restored.

[@nickmcmahon1568]

My grandfather worked for Control Data and traveled the world as a Technician to repair Control Data`s Supercomputers like these, it amazes me how far advanced we are now compared to these for Supercomputers of today.

[@twol78s90]

I did systems operations on Tektronix' Cyber 73 from 1977 through around 1980. I really loved that machine. Those console displays were awesome. I once had the console display system literally blow up, shooting sparks and bits of molten metal out from beneath the tubes toward me. Fortunately, I was sitting in a chair with wheels. Just before the fireworks started, I noticed one of the displays shrink down to a bright horizontal line, and heard a loud squealing noise, and pushed myself away from the console just before it blew up. The Cyber kept running, but the console was toast. It took two of CDC's Customer Engineers about 6 hours to fix the console. One of Tek's system programmers had written a program that could mimic the console displays on a serial terminal connected to the system, so we were able to maintain operations despite the console being down. Many great memories of those days. Very sad that COVID-19 has shuttered LCM+L. I hope it returns some day.

[@VulcanOnWheels]

3:33 Monica wants to ask a question, gets interrupted, and then she doesn't appear to get another chance to be heard. That would be a shame.

[@carlblaskowitz7817]

There aren't that many girls interested in vintage supercomputer history... if you find one... keep it!

[@jvolstad]

Who is the girl?

Any girl interested in old mainframes is someone I'd fucking marry

[@n0tyham]

I went through Control Data Institute in 1975, where we worked on an old 3300 computer. It used 1604 logic circuitry with just a few transistors on each board. I used to put in Field Change Orders on the machine, and freeze my fingers off because of the 40° chilled air blowing on the backplane! This video really brings back some memories. BTW, if you watch "Colossus: The Forbin Project", you'll see that Control Data Corporation loaned the studio a lot of computer cabinets for the movie.

[@pentogram23]

I worked at imperial college in the early 1970.s using 6600 6400 and cdc 17oo computers. Steve Page was my group leader and Marcel Ottoy the manager...great place to work..

[@n0tyham]

BTW, I started programming on an IBM 1401 when I was 12 years old in 1968. I had already taken a course from Cleveland Institute of Electronics by mail, which was how I got my start in radio and computers. It's too bad I never worked on a 6600 system. I did get to use the CDC Star computer when I went through Control Data Institute, and they were using Plato plasma display terminals.

[@novaseline4u]

I worked on these machines with CDC 40 years ago, at Picatinny Arsenal in Dover NJ. We had a 6500 and a 6600 there. We also had an Extended Core Storage unit on the 6500, The 6500 was later replaced with a used Cyber 74. An 808 disk drive was there. 405 card reader, like in this video. This 406 is newer than the 65oo -- it is painted in Cyber 170 paint scheme. We had a bunch of 833 disk drives, and 66X tape drives. The reason you see those deep wiring mats in the mainframe is for signal timing. Timing was determined by the length of wire the signal had to travel through. Biggest issues I remember with the mainframes were the memory stacks, and the refrigeration units. These machines also ran on 400 Hz power, like aircraft electronics do. We had motor-generator units that provided the power to run these machines.

[@kwmcc]

2:36 ... what type of problems were addressed on this machine? He mentions some cool research. But what I remember was CS 220 Introduction to Fortran. Using punch cards. That's the actual computer I used to learn programming, fall of 1978. It was in the first basement of the Math Sciences building. The second basement was the card punches, card readers, and printers. The next year we were programming in C on PDP11/70s and VAX systems in the EE department using terminals.

[@robertmaclean7070]

Sounds like Monash University Computer Centre? Where I worked for 14 years.

[@ComputerHistoryArchivesProject]

A great museum and a great collection of vintage computers! Thanks for posting this!

[@howardwhite9773]

Did I hear Lath say that this CDC 6500 was at Purdue University?? I used that one!! :) Compass is an elegant assembler language.

[@DaiyuHurst]

Yes, it was I used it there in the Summer of 1974. Then helped debug it, and provided the operating system software.

[@computernerd8050]

I love COMPASS! Used this machine's "brother", a CDC 6500 at Michigan State University, while I was a student. They later added a Cyber 170-750... a newer, faster machine with semiconductor circuits and memory, unlike the 6500 core memory. The Cyber 750 was mostly code-compatible for users of the earlier 6000 series. Their 6500 ran along side the newer 750 for several years. I miss the CDC machines!! After college, I did work at 2 places with CDC computers. I miss them all!

[@michaelbauers8800]

I actually remember Compass from being in Minnesota in the 1970s and 1980s, and have some access to a Cyber machine at the University of Minnesota.

[@mendax1773]

Indeed! Compass was an awesome macro assembler. And I still have the book for the instruction set for the CDC 6000 series machines (and its children and grandchildren, the Cyber 70, 170, and 180 (if properly microcoded) computers. The most impressive code example in that book is the routine that saves all the register contents to memory, not simple task programmatically given that you must, for example, put the address in memory to write in A7 in order to store the contents of X7, which means you have to somehow save the contents of A7. I think this was the source code to the routine the OS used to do exactly this when generating a register and core dump when some CPU exception, like an illegal memory access or a divide by zero occurred. I remember reading somewhere that it was written by the legendary Greg Mansfield. His fingerprints could be found all over CDC operating systems.

[@marcos.666]

@@mendax1773 Exchange Package. I'm pretty sure that the exchange jumps where the exchange package was loaded from central memory into the registers or the registers were dumped into the exchange package in CM were by instruction, either CEJ when the CPU was running CPUMTR or by MEJ from PPMTR running in PP0. The interaction between CPUMTR and PPMTR was quite elegant as was the CPU/PPU architecture.

[@RME76048]

At 4:10 he asks, was this the only kind of volatile memory that the machines had? The neat thing about core memory is that it is non-volatile! When you write a 1 or 0 to a core bit, it stays. To read a bit required destroying the recorded state, which then required rewriting the bit after a read, thus slowing reads down. An historical note: The Challenger space shuttle (and as I recall, all of the shuttles) used core memory encapsulated in a block of acrylic, designed to survive orbital re-entry. That proved crucial in 'seeing' what the orbiter's software was doing moments before its tragic end.

[@eddiehimself]

When I did my engineering degree, they literally showed a video of Roger Boisjoly as a way of telling us not to let people put financial/time pressures over critical safety measures.

[@rdvqc]

What was even more cool, you could "deadstart" Kronos (NOS) from memory in a couple of seconds.

[@gmcenroe]

I am reading The Supermen The Story ofSeymour Cray and Technical Wizards Behind the Supercomputer, nice to see this surviving computer.

[@PauldeSwardt]

I started my IT career at Control Data - East Barnet - London - UK in 1980 - currently with IBM in CZ - Remember the Snoopy screen saver!

[@herauthon]

so WREN rings a bell ?

[@fromanabe8639]

I was employed at Control Data Arden Hills and Plymouth plants from 1969-76. I was sent to Seoul, Korea in 1971-72 where I coded modifications to the Cyber operating system that allowed printing Korean (Hangul) on a CDC line printer equipped with a special print train.

[@rogert9754]

I worked at Control Data, Taiwan. In 1988, I had been to ETA, "Energy Park" @ St. Paul, to be trained as a Software Engineer for ETA 10. During that time I had been to Arden Hills a couple of times. It's a shame, in less than one year, the ETA was discontinued. BTW, Control Data, Taiwan also had a "Power Distribution" project for Taiwan Power Corp. which had several Taiwan Employee working at Plymouth Facility. Remember been to that facility once or twice. I spent around 3 months in Twin Cities, from 105F hot summer, fall foliage, to snowstorm the day I left, interesting experience.

[@leodf1]

Good for you

[@robertsrodulski9438]

CDC 3100, 8490 and 160-A operator here, 1971-1974

[@milo9942]

Wow! Super cool! I've always wanted to use one of these machines!

[@your_local_dummy4137]

Started my computer life on a CDC 6500 as the first year student. Of course first year students were always stuck with punched cards and typing in the card punch center. That was a noisy place. The computing center to house the 6500 was the floor area of house. By Honors in Physics I had access to the Physics terminal room and could run my own stuff and use a VDU, revolutionary. I remember late nights playing an early computer game called star trek, of course poor graphics but it was fun. Those old Fortran physics modelling and scattering programs lasted for many decades. Why rewrite good well proven code that works and only goes faster and faster on modern emulators. Great look back, good memories and thanks.

[@lawrencedoliveiro9104]

2:23 The 6600 would have been the first. I think it was about 50 times faster than anything from IBM. Which didn’t stop them from promising the amazing new 360/90 machine, that would retake the top computing crown, just you wait and see. So customers held off from buying CDC, until the IBM machine actually appeared, and completely failed to live up to expectations. CDC sued over the FUD campaign. And wonder of wonders, they won. Worth reading: “The Supermen” by Charles J Murray.

[@movax20h]

50 times faster. Not really. It was maybe 3 times faster than IBM machines, but IBM machines were release few years earlier. Still 6600 (and 7600) was a huge achievement, and a had a lot of pioneering solutions, that become standard only later.

[@GH-oi2jf]

I used these CDC machines. We never called the 6500 a “supercomputer.” The 6600 was called that when it came out, but the 6400 and 6500 were compatible, but simpler and slower processors.

[@paulcarter7445]

Great video. Small correction - core memory is not volatile - it retains all its data when turned off. Reading does destroy cores' content and requires a write-back, but that doesn't make it volatile.

[@GH-oi2jf]

Correct. Bonehead mistake.

[@jcupp2]

I learned to program on a CDC6600 at IU in 1980. I never actually saw the computer I was using as it was locked in the basement of the computing center. I just used one of the remote I/O nodes on campus but I did get to run my own cards. There was no sorter available and they never changed the ribbons in the key punch machines so sorting a deck was always a good time. But I learned to just 'read' the cards.

[@RME76048]

WOW! A functional CDC 6500 Supercomputer! And you can run your own code on it too? **Amazing!!!** I have no idea when I might make it [back] to Seattle, but when I am there I'll be sure to visit and find out more!! Keep up the GREAT WORK!

[@DaiyuHurst]

You can log into it remotely.

[@tickertape1]

I do quite often very fun machine to work on but it’s up time is variable

[@GH-oi2jf]

The 6500 was never a supercomputer.

[@simsym3293]

CDC 6500 Computer Operator from 1974 till 1979 here!

[@jeffpiepmeier9411]

Sim Sym when did that CDC arrive at Purdue? My dad was systems analyst for Purdue’s CDC circa ‘70-‘71.

[@douro20]

I wonder if there are any complete 7600s?

[@jayc2469]

Gotta luuuurve Monica! She makes a refreshing change from the old cronies in this field! In fact, I shall subscribe and thumbs up too, for good measure!

[@VulcanOnWheels]

Do you know if she graces other KZbin videos?

[@jamieboulton1705]

I wanted to know what question she wanted to ask at 3:32.

[@dainipeagram4837]

Great video, but OMG, let her ask her questions!!

[@rdvqc]

Some of the most fun systems I ever worked on. I worked for Control Data Canada in the 70's and later with Pratt & Whitney Canada on 6000, 7000 and Cyber 170 systems under all the various operating systems. BTW, only students typically kept their programs on cards.

[@lawrencedoliveiro9104]

Hmm, the 6500 seems like a later model (not mentioned in the 1965 hardware reference manual, which only lists the 6400, 6600 and 6800). Looks like a dual-processor version of the 6400.

[@DaiyuHurst]

That's exactly what it is.

[@robertmaclean7070]

Thats right. A dual processor version of the CDC 6600. Seymour Cray didnt like multi processors.

[@lawrencedoliveiro9104]

@@robertmaclean7070 But the Cray-2 had 4 processors.

[@GH-oi2jf]

@@robertmaclean7070 - No, a dual-processor version of the 6400.

[@GH-oi2jf]

@@robertmaclean7070- Cray had no reluctance to use a few processors, but the 6600 was too bulky to fit two into one system.of four bays.

[@tonyzambos3138]

Worked on 6600, Sn #9, and dabbled a little on a 6400. A 6500 is the combination of a parallel processor, 6600, and a sequential processor, a 6400. And just for laughs, that is an oscilloscope sitting on the floor next to that 405 card reader, have one in my basement. We trouble shot problems down to a transistors and replaced them. Timing in these machines was done by tuning the wire lengths between modules. A nano second is seven inches of a twisted pair of 24 gauge. Wished I've closer. It be great to one of these old machines.

[@GH-oi2jf]

Wrong. A 6500 has two processors of the 6400 type. There may have been a 6700 with one processor of each type.

[@tonyzambos3138]

@@GH-oi2jf Yes, you are absolutely correct, my mistake.

[@douro20]

The high-speed card reader looks like a Documation product.

[@robertmaclean7070]

The card reader was made by Control Data. Documation made Line printers in Melbourne Florida.

[@lawrencedoliveiro9104]

2:11 That console says “6600”.

[@lawrencedoliveiro9104]

2:20 Ah, the other one says “6500”.

[@gerardooviedo4145]

No el equipo más rápido, sino el más político

[@jayc2469]

There's a Chiller on the Roof!🧛‍♂️.♫. A chiller on the roof again!😅

[@rodrigonegrelli8668]

Crysis wheel?

[@marcos.666]

I see Snoopy. The 405 operator failed to hit the "EOF" button below the "Read Cards" button to clear the buffer. The sequence was Motor Power, EOF, and Read Cards. kzbin.info/www/bejne/rYWVpmVjpp1rn5o

[@OldDogNewTrick]

Back when computers were built by IBM and the BUNCH - Borroughs, Univac, NCR, Control Data and Honeywell.

[@MaquinasAbiertas]

Is she ornamental? Why didn't she ask any question?

[@jvolstad]

I would submit RPG II programs on punch cards. Good times.

[@gerardooviedo4145]

Esa máquina era más poderosa que la mejor IBM, pero sobrevivo, no el más

[@movax20h]

A bit of exaggeration about a phone being 2 million times faster than CDC 6500. I would say 10 thousand times faster is a more accurate measure. Even today (this video was 7 years ago), it will not be millions. You would need a workstation class CPU with many cores, or few GPUs to achieve 2 million times faster mark. The supercomputers like this were pretty fast for the time, and in fact minicomputers and PCs decade later were way slower, so this is why we think about this million time faster figures. But performance per watt, definitively over million times better.

[@kevinhoward9593]

that super computer is probably 250,000 less powerful then my iPhone 8.

[@ingusmant]

Theseus' computer

[@iptashR]

0:32 show what he holds, not a girl

[@bloodyhell6378]

Mmmmmm dohhhnut...

[@stevebez2767]

I recall those guys determed too build that an fully personnel what was assumed a 'Seymour cray'!! This guy in footage pursued them initially suggesting many embedded says of common things thought by himself as later followed even had too talk him out of suing for copywrited etc!!! As no one would put direct things he began swearing he had programmedPCs too do which was also a conundrum thought clever but fact remains boast would accuse and stand factor that is followed of unsound says to belief of professors which CDc kind of idol bypassed many times perverting manner of ship along thought word terror baited reams of subject invents thought too be online computer info etc,d.n.a,atoms,scientific study as simply unsaid,Unknown,quest of whom do you debate this with then,crazy key line cipher fact of having said non sound use of word too invent such too a do,order of looking cinq port sailors running red robe crash test dummy's for law rule Lord robot loony car lent tree really Jon God a new mode tar jolly Jack roundabout bill life's of but its me whole world you crazy fuckers,chase of snoop fuckers oooo!??

[@carlblaskowitz7817]

And increase the dosage!