1950-1983 Computer History at Lawrence Livermore Nat's Labs, UNIVAC LARC, IBM, CDC, CRAY

[@ChatGPT1111]

Dr. Teller came to my physics class to lecture one week at Whittier College in ‘79. It was very enjoyable but we were all humbled by his intellect.

[@ComputerHistoryArchivesProject]

Hi Melodius, that sounds great! I would have loved to hear Dr. Teller speak live. Never boring, always enlightening! You were lucky to have that opportunity. : ) CHAP

[@scottcass4243]

I started my career in IT in 1980 & I recognize many of the later peripherals. I don't miss standing at a card reader for hours at a time.

[@ComputerHistoryArchivesProject]

Hi Scott, I can relate to that too. Lots of wait time to use a punch card machine or computer CPU time...Like the old army slogan "hurry up and wait." Thanks for visiting our channel! ~ Victor, CHAP

[@jaydrew4787]

The term "Concentrator" that John mentions at the 16:14 timestamp was an in house term we used which the modern vernacular would call a terminal server. The in house designed teletype multiplexer had 128 teletypes attached to it, each running 110 baud. The hardware of the "concentrator" would assemble the serial bits of the communication (before the days of the UART) and present them to the PDP-8 which via an in-house designed link would send the parallel data to a mainframe (6600 or 7600). When a character was sent from the mainframe back to the user terminal, the parallel byte would arrive at the PDP-8 and it would present it to the concentrator which serialized the data to the appropriate terminal. The next generation of a TTY Multiplexer was designed to handle 256 terminals running up to 9600 baud. The controlling computer for that system was a PDP-11. The line cards had UARTs on them and the serial to parallel was done on those chips. The third and last generation of the in-house terminal server was called the Micro-mux. It had 128 line cards each with an Intel 8039 processor and the processing load was taken off of the central processor which was a DEC LSI-11. The PDP-8 and PDP-11 systems were attached to the mainframes with in-house designed interfaces. The Micromux was attached via the NSC Hyperchannel. John programmed all three systems and did double duty on the Micromux, he wrote the codes for both the LSI-11 central computer and the 8039 microprocessors.

[@ComputerHistoryArchivesProject]

Hi Jay, Thanks very much for your detailed info! It sounds like you have a lot of tech experience in this area. ~ Victor, CHAP

[@jaydrew4787]

@@ComputerHistoryArchivesProject I worked at the Lab's computer center from 1974-2013. I was hardware support and worked with John from 1975 to when he retired in 1993.

[@cyberyann]

Thanks for the clarification!

[@programaths]

I like also the "directed graph" appellation for what we know as tree directory now. Though, it also tells that we could have many roots (sources).

[@ulyssespascal4875]

I came across the term "concentrator" when researching NASDAQ's original network in 1971. Do you happen to know if the NASDAQ concentrators were similar?

[@tubastuff]

An old CDC STAR guy here. I recall that LLNL folks had bootleg programs that would access all one's files to keep them from being "photostored". That was a pejorative term, since one was never assured of getting the files back...

[@ComputerHistoryArchivesProject]

Hi Tubastuff, that is an interesting bit of LLNL history. Didn't know that! Did you work at the Lab yourself? ~ Victor

[@tubastuff]

@@ComputerHistoryArchivesProject Hi Victor--no, I was part of the STAR group at CDC Sunnyvale Ops (SVLOPS). STAR hardware development was handled out of Arden Hills, but SVLOPS handled almost all of the software effort for a time. We had two STAR-1B emulators at SVLOPS, but they were incredibly slow and flaky. Some of the group came from the Canadian STAR development operation, which used a STAR-65, but that was junked when that operation was closed. So if we (in the systems group) had to do anything serious, it was either grab a flight to the Twin Cities (we had a department parka for the winter) or finagle some CE time at LLNL. Eventually we set up a remote console link to the STAR at Arden Hills, but it was late in the project. Out of all of this the ETA 10 grew.

[@jaydrew4787]

There was a time limit for files stored on the disks attached to the mainframes. This was done to reclaim space so users could run their codes. The "bootleg" programs were scripts that users had that would "touch" the files thus updating their timestamp. Files did not migrate to remote storage like the CDC Mass Storage system or the IBM Photostore without the user placing the file onto the devices themselves. The CDC Mass Storage had less space but quicker access, files written there had something like a two month limit before being erased. The Photostore was the long a long term storage device. The statement of "never assured of getting the files back" is incorrect. The Photostore at Livermore was pretty reliable, mainly because we had a superb on-site IBM engineer. The other systems at other locations did not have that fellow and their systems did not fare as well. The biggest problem with the Photostore was the very slow speed. The system would fill up and the older files were taken out of the machine and put on a shelf. If you requested your old file and it was not in the system, operations staff would have to retrieve the cell containing the strip(s) your file was written too, and that cell would be manually inserted into the machine. All of that took time, sometimes several hours. As to the "automated" file ticklers, I believe the operations staff had discussion with the users and their use diminished. Eventually long term storage returned to the world of magnetic tape in different formats. But if you ever wanted to see a Rube Goldberg machine in real life that WORKED, it was the Photostore.

[@pereimar]

Amazing!!! I had no contact with these machines, as I was born in 1971 (Rio de Janeiro - Brazil) and started working in IT in the 90's (1998 precisely). But I am fascinated by old mainframes and computers, their history, functioning and legacy! Especially the DEC PDP-7 and PDP-11, where UNIX was created/developed. Fortunately, I had the opportunity to meet one of his descendants... FreeBSD 4.7-RELEASE! I have it installed until today on my Compaq Prosignia 500 (Pentium 90Mhz). I learned a lot from them! Thanks God! Excellent and educational video! Congratulations!!!

[@ComputerHistoryArchivesProject]

Hi Marcos, thank you very much for your excellent feedback and interest! I understand how you can be fascinated by the old mainframes and other computers, there is a fascinating amount of early technology. Later machines and desktops are also great tools to explore with. Thank you again. Hope you will explore our channel further! ~ Victor, CHAP

[@pereimar]

@@ComputerHistoryArchivesProject Thank you Victor! I will certainly enjoy and explore all the videos on this channel! I'm already doing it! The history of these great computers and their creators is fascinating! Allow me to ask a question: Is it still possible to find a fully functional IBM 360 mainframe or similar in any museum in the USA? It would be awesome to watch a demo from power on, boot process, program loading and execution, data input and output, etc! PS: I'm using a translator (Thanks Google Translate). If there's any word out of context or improperly translated, I'm sorry! LOL

[@ComputerHistoryArchivesProject]

I think Curious Marc has some great info on his channel... and check out the Computer History Museum, and the Living Computers Museum as well. Lots of folks have put great time into restoring some of the vintage giant machines! ~

[@bloqk16]

As a youngster in the 1950s, I recall when the LLNL was locally known as the Lawrence-Livermore Radiation Labs; as back then radiation, at least with the public perception, was considered a lot more benign than it actually was. I find it amusing that the place changed its name to sound less threatening.

[@vm2463]

30 years into the future for him is 10 years ago for us...

[@videosuperhighway7655]

One day in the future people will have super computers that fit in a pocket and come with a built in camera and high speed FFT processors so the user can record cats playing and post this video on a large scale distributed network.

[@AtulShetty]

Why would someone dislike this awesome video? It is beyond my comprehension

[@cosmicraysshotsintothelight]

It is awesome. I also liked seeing some of the computers at the National Cryptologic Museum. I find it funny that my 6 year old smart phone is more powerful that the original Cray was/is. And WAY less Watts!

[@richardgray8593]

Maybe people who don't like to see tax money wasted.

[@josugambee3701]

@@richardgray8593 "Wasted"

[@kathleenyurick3902]

It's what they do.

[@techwithdave]

I’ve literally seen the thumbs down getting highlighted despite the fact I clicked the thumbs up button.

[@MicrobyteAlan]

The CDC 6600 was the 1st computer I serviced. Then went to DEC to service DECsystems and VAXs. Interesting video and well presented. Thanks

[@ComputerHistoryArchivesProject]

Thank you. I only wish we could find an original film, in closer to HD quality. : )

[@tonyzambos3138]

If Livermore only had one 6600, then it was the first computer I serviced when it was sent to the Fermi Lab.

[@osvaldocristo]

Another great memory this video brings me: in the very beginning of 1980s our university received a CDC 6600 or 7700 (I cannot remember the exact model) to pair our B6700/6800 as our main computers. I was a graduate Engineering student. I had a few projects I used both computers. Thank you to brought a such memories.

[@ComputerHistoryArchivesProject]

Hi Osvaldo, Wow, the CDC 6600, that was a famous system! I search for films about that machine, but they are very rare. I think I have one or two short clips of it on the channel here. We have some on the B 6700 series too. You have an amazing background! ~ Victor,CHAP

[@bobelliott2748]

Lol. I did a liberal arts degree with a major in English. But, it being a liberal arts degree, I also did science courses. I remember modeling the motion of a pendulem using a teletype machine for input and the output. It worked! It used a forest of trees to print the output but I got an "A" on that lab tutorial. I believe that my university in 1974 had two computers that occupied the entire basement of the mathematical science building. I enjoyed this video

[@cookie7789]

Thank you for this video!!!!

[@ComputerHistoryArchivesProject]

Hi cookiefortress, I am very glad you enjoyed this. Thank you for the feedback. ~ Victor

[@8BitNaptime]

Is there a Windows 10 driver for that laser printer? I want one!

[@S_M_360]

Man, that octopus network sounds badass 😍

[@DrB1900]

It's amazing that today you could buy a $200 computer that would greatly outclass that Cray X-MP.

[@captainkeyboard1007]

This show is very nice. I wish I was in that Computer Museum.

[@DASDmiser]

11:18 CDC mass storage device, sold in the IBM market place as a 38500 (IBM had helical scan carts model 3850). My first job out of college was coding "drivers" and support routines for the 38500 in IBM 360/370 market (across all versions of MFT, MVT, SVS and MVS operating system). The system automatically managed a hierarchical storage dataset usage on disk with of a tape repository on a data set by data set allocation basis. Pick to empty cell error advised operator to look on the floor. Besides Seymour's developments in vector registers and heat management Cray disk drives enjoyed much higher bandwidth by enabled for parallel read/write across multiple platters by simply employing wider channels and enabling the multiple heads in simultaneously. A novel idea that was never adopted in the larger market place.

[@ComputerHistoryArchivesProject]

Hi Dean, thank you very much for sharing that info about the IBM38500 and your experience with it. It is great historical info that one does not usually come by, except from people with first-hand experience like yourself. Very cool. ~ Victor, at CHAP

[@DASDmiser]

@@ComputerHistoryArchivesProject I should be clearer, it was the CDC 38500 MSS, marketed to IBM customers as the 38500. It competed against the IBM 3850 MSS but with many distinctive physical and operational differences. Both MSS systems employed cartridge tapes as repositories for a storage hierarchy and transparently staged the user's data set to DASD (IBM on a cylinder basis, CDC on a dataset basis).

[@ComputerHistoryArchivesProject]

Hi Dean, thank you for the additional explanation. : )

[@tubastuff]

@@DASDmiser My recollection when seeing the 3850 at an NCC was that it was incredibly complex mechanically (cables, gears and pulleys). I asked why there were *two* picker arms and was told that it was designed that way in case one arm failed. The CDC unit was far less mechanical, IIRC. I called them "beer cans in a beehive"...

[@zatozatoichi7920]

The main unit of the Cray XMP is beautiful.

[@ComputerHistoryArchivesProject]

Hi Zato, agreed! The simple yet powerful Cray design is iconic! ~ Hunter, at CHAP

[@thesteelrodent1796]

it looked great, but it used an immense amount of power and just a few years later it was hopelessly outdated

[@GH-oi2jf]

This video shows a Cray-1, not an XMP.

[@Taigan_HSE]

@9:46. That’s the same capacity I’ve got on my phone.

[@mmaranta785]

Wonder how many paper jams that printer had

[@raphaelmsx]

Fantastic video!!! What's the name of the 9-track tape library at 9:45? Never heard of a tape library on open reel tapes until now!

[@ComputerHistoryArchivesProject]

Great question! - After some research, I believe this device may have been a version of the “Braegan ATL.” “Braegan Automated Tape Library (originally from mid 1970’s) consisted of an enclosure housing several thousand mag tape reels and a robotic mechanism for moving reels from shelf storage to self-threading tape drives. Originally developed by Xytecs, sold to Calcomp, then to Braegan, then acquired by Digital Storage Systems of Longmont, Colorado.…” Source Publication: "Advances in Computers," M.C. Yovits, Purdue School of Science, Vol 27, Academic press, 1988. Google Books has a preview… -- Thanks for asking! and thanks for watching! ~ Victor, Computer History Archives Project

[@raphaelmsx]

Hi again! I found three more pages regarding this system (the 9-track ATL), two ads and one review:

[@ComputerHistoryArchivesProject]

Those are some fascinating references! Thanks!

[@ComputerHistoryArchivesProject]

Excellent! Good info those reference links. Thanks for taking the time to do that. Victor, CHAP.

[@muhammadramardhikasofyan3763]

Wow.. 10 characters per second

[@markarca6360]

10 bytes per second (80 bits per second)

[@jaydrew4787]

10 characters per second is correct. The PDP-8 Terminal system ran at 110 baud. A transmitted serial character consisted of 1 start bit, 7 or 8 data bits, a stop bit and then a parity bit. With an 8 bit character would be made up of 11 bits total. Divide the baud (110) by the bits (11) and you get 10 characters per second. All of the terminals of the PDP-8 systems ran at 110 baud. The PDP-11 systems had various baud rates running from 300 to 9600 baud. The Micromux system was the most flexible, each line card could run at its own defined speed. We ran current loop circuits for the 110 and 300 baud systems, RS-422 for the higher speeds. We were pretty much limited to 9600 baud due to the type of wiring and the distance we had between buildings.

[@lidarman2]

I think it is funny in the Univac photo they showed; It comes with an oscilloscope. Prolly a Tektronix 585.

[@frankowalker4662]

The perfect Sunday! Nice weather, a few beers and wonderful documentaries.

[@ComputerHistoryArchivesProject]

Many thanks! ~ Victor, at CHAP

[@skeletor8250]

Wow!, I'd like to have that Cray computer . I'd be able to BLAZE through multi-digit additions with that !

[@mikakorhonen5715]

How would you power that monster?

[@ChatGPT1111]

I heard that they upgraded it to actual two digit multiplication. It only required 22 jigawatts of power.

[@areality40]

I like the music in this video. Kind of reminds me of what I am currently making.

[@Parakinese]

Awesome!🥸

[@punman5392]

This is basically ground zero for the internet.

[@ironchef3500]

Neat

[@majikglustik9704]

Cool vid! 📀__⚠️__📀

[@MarkWhich]

capacity of 2 Trillion Bits?

[@jaimdiojtar]

32mb of ram for that beast

[@michaelodendaal5776]

Wow 2t-bites on tape

[@lawrencedoliveiro9104]

Regarding the copyright restrictions, wasn’t this produced at a Federal agency funded by the US taxpayer? Which means it is in the public domain?

[@ashz2505]

Not everything funded by any county's tax payers are public domain (for that country citizens). Actually very little of tax payer funder ANYTHING become public domain. I do not believe this needs to be explained.