Before IBM was synonymous with personal computers, they were synonymous with large computers. If you didn’t live it, it was hard to realize just how ubiquitous IBM computers were in most industries. And the flagship of the mainframe world was the IBM System/360. For a whole generation that grew up in the late 1960s and early 1970s, a 360 was probably what you thought of when someone said computer. [Computer History Archive Project] has a loving recollection of the machine with a lot of beautiful footage from places like NASA and IBM itself. You can see the video below.
Not only was the 360 physically imposing, but it had lots of lights, switches, and dials that appealed to the nerdiest of us. The machines were usually loud, too, with a Selectric terminal, card punches and readers, noisy 9-track tape drives, and a line printer or two.
While even a supercomputer from the 1960s doesn’t seem very powerful today, the 360 holds up pretty well. Most of them were 32-bit machines although there was at least one 16-bit bargain model and one that did everything 8-bits at a time but still made you think it was churning out about 30,000 instructions per second. You could even join some of them together to get more power, although memory bus contention made that less effective than you might think. Some of the higher-end models used 64-bit memory, parallel execution, and virtual memory.
One hallmark of this family of computers is that it had sophisticated I/O channels that could interface a lot of devices to the CPU. Of course, you needed a lot of devices when a hard drive cost more than your house and would hold 5 or 10 megabytes or so. Memory ran from about 8K to 8M — quite the range. Usuallly, the computers were leased not bought, so price comparison is hard. But a very large System/360 Model 195 was quite fast, could multiprocess, and had a whopping 4 MB of memory. The cost? Somewhere between $7 and $25 million in 1971 dollars!
Computing was a different world back then. Raised floors and special air conditioners were the order of the day. We also probably should have been wearing hearing protection! Amazingly, as common as these were, not many of them have survived and many of those aren’t working. These days, DASD and IPL aren’t common terms in the computer business, but when the 360 ruled the data center, they were terms you heard all the time.
If you are lucky, maybe you’ll find one stuck in a barn one day. It could happen. Just try to make sure the barn is close to your house.
I worked on an IBM 390 (which had multiple 360s as ancillary processors) back in the 1980s. COBOL, BALR (assembly), C, and JCL (Job Control Language).
In Dungeons and Dragons fantasy world terms it was a ‘Legendary Beast’. The whole thing filled a room bigger than most high school gymnasiums – not counting the tape storage vault.
BAL is the assembly language, which generally started programs with BALR – branch and link based on contents of a register
JCL…eyeball starts to twitch…
Javascript on the server is not the worst pigf@!k in computer history. Holds a solid lock #2 though.
“… a hard drive cost more than your house and would hold 5 or 10 megabytes or so.”
Sure, but that didn’t matter a ton because all of the data (and some of the programs) was on punchcards and/or magnetic tape. :-)
Hard drive? Mag tape, up to 8 drives or a box of cards.
No FASTRANDs? Also guys/gals, Fixed Disk Drive, please! :)
IBM 360 was my first exposure to computers back in the 70s in college, on which I learned assembler and Fortran. Stacks of Hollerith cards made on 029 punch card machines. Such memories.
Sounds Really familiar. My location was the IBM lab in Böblingen Schönaich 1970-1973
I worked there 1970-1981 in the computing center
Luxury!
When I was a kid, we had to make our own punched cards out of taped together discarded packaging.
We dreamed of having a card punch, made holes in our ‘punched cards’ with pocket knives.
Then we had to wait for months for printouts, to find out we had a syntax error.
Of course this was in Yorkshire, so we were happy, thought it was great.
Basically the same for me but we had to sharpen rocks by grinding them together because we didn’t have pocket knives lol.
We were spoilt, we had the cards that had the rectangles mostly cut so you use a pen to push it out and after complaining about that we got a mark sense reader that needed re-calibration every two cards lol. I sense some punishment in that. We had to manually feed cards so we used fan form and a break meant you made a mistake and removed that card. Fan form cards had a very marked advantage when you dropped a stack.
The author here said these areas were noisy but some were near silent. Selectrics weren’t used on-site much. Almost everything else had an acoustic chamber. You couldn’t possibly sit in the same room as an (actual) line printer unless it had an acoustic chamber.
Actual line printers were bolted to a concrete floor so the floor would move at exactly the same rate.
My wife’s aunt learned COBOL on a 360 many decades ago. Still well used in finance. Probably an emulation of an emulation of an emulation of the 360 at the base level.
!**! It’s Emulators! All the way down!
I started on an IBM 1620. If you planned to do a multiply, you had to preload the multiplication tables from a punch card deck. Interesting note- the IBM System/360 operating system was hailed as one of tech’s great achievements. Turns out the code that actually shipped with the machine was written by a small IBM group in Europe who were responsible for the IBM 1401 computer/peripheral controller.
I supported both the 1620 and 1401 as an IBM systems engineer… One was a scientific oriented machine and the other commercial… They were my introduction to computers, then the 7080 and onto the 360 family…
Worked as operator on 360’s in 1970 running 1620 emulation. Had a “halt” book for each program and read the console lights to get halt address and dial in address to branch to. Fun times. Later Prudential trained this barely high school graduate as a COBOL programmer. Had a full career in IT on many platforms.
In 1974/75, as a high school senior and grad I had a part time and summer job in the pump engineering department of the then Allis Chalmers, now Siemens plant in Norwood Ohio. They had a 360 downstairs which I never got close to, and an 1130 in my department with an attached printer that I loaded with paper a couple times. Don’t remember now what they actually used either computer for. Spent a lot of time with an electric eraser and pencil making changes to specs on transparencies, and going through flat file drawers looking for misfiled drawings. Before CAD, engineers/draftsmen sitting at tables drawing on paper or transparencies!
I love OS/360 Assembler. I wrote a bubble sort program using it.
I was fortunate enough to start with punch cards and paper tape, progressing on to the Teletype, and then the coveted CRT terminals, all in the space of two years.
Learned Pascal on a 360/44: already knew FORTRAN. After you punched your cards, you’d rubber band them and put them .. into an input location. I think it was a window from the machine room into the keypunch room, and it had a broad sill. The operator would open the window and get your deck. If it was a new deck, it would likely come back quickly, into an alphabetized pigeonhole with a door on the back that made a particular sound when opened and closed and, allowing wash of white noise into the keypunch room. When things were busy, there was a lot of waiting.
Your deck was returned wrapped in a green bar printout with your account name on it, and a listing if you requested it with JCL, which you pretty much always did if you were starting a new program. You’d fix your syntax errors, check your JCL cards again, and resubmit it.
There was always a computer operator on duty, day or night. Printout pages, cards read and compute time all had costs in quatloos. I think a student job could run for as long as two minutes. Punched cards were available in the bookstore.
I got my hands on an IBM System 23 all in one at a garage sale. The keyboard, monitor, and dual 8-inch real floppy drives. It also came with a bunch of the old floppies that you booted it from. It was the “Brady” operating system. Those were the floppies that were just the magnetic film with no package like the 5.5-inch.
I got it home, set it up. Put the floppies into the drives, and it booted up! Played with it awhile and had some fun. I lived in an apartment then (now own my own house). Someone broke in a stole it.
I decided then to buy my own home!
Ahhhh the memories! My dad worked at IBM from the 50’s through the 80’s as a repair technician and then the head of their management development training program in White Plains, NY. I still have a number of repair parts including silver contact relays and some vacuum tube assemblies. In middle school I made a 10-bit binary counter from the relays as a timer for my science fair experiments on hydrogen-oxygen fuel cells. The device is currently sitting in my electronics project room and still works!
We had an IBM 1130 at my high school which I used to learn FORTRAN. The first program I wrote was for converting Fahrenheit to Celsius temperature readings. We had to create the program using punch cards and feed them into the reader. If you were lucky, the program would work and there would be a printout on the console printer.
I also took an APL programming course using a remote terminal dialed into a local 360 timeshare system. It was slow, but it eliminated the need to use punch cards. My high school science teacher had a portable timeshare terminal that he would let me use. I also remember going to different local college libraries to use their terminal.
I went to college at RPI in Troy, NY majoring in physics and math, graduated in 1978. We had an IBM 360 that was fully configured with every peripheral IBM had at the time. It was so big and generated so much heat, it was used to heat the building where it was housed during the winter! I created very large programs for my classes using FORTRAN that sometimes had 3 to 4 boxes of punch cards. Every day I would drop off the latest version of my program at the computer center and return the following day to obtain my printout. It was such a slow and arduous process debugging my programs!
As a senior in college I finally got a job at the computer center diagnosing, cutting, and splicing 9-track tapes. I typically worked the midnight shift because it was easier to do my work with no one else around other than the system operator.
After graduating college I joined the Navy as an officer and had various onboard and onshore assignments. My last assignment (1982-1983) was at the Navy Regional Data Automation Center (NARDAC) in Norfolk, VA where I worked for Grace Hopper. Grace was one of the pioneers of the computer industry who worked on the Harvard Mark I computer and created the FLOW-MATIC programming language which was the foundation of COBOL. I led the Microcomputer Technical Group which reviewed the latest microcomputers for potential use by the Navy. Because Grace led our program, we were able to get our hands on the latest microcomputers and software. We also started a newsletter called Chips which focuses on the use of information technology within the Navy which is still in publication today (40 years!).
I joined IBM in 1965 a week out of high school. It was the age of the S/360. The company was flourishing and we got everything we needed. Great time back then!
someone i have read about started a job , the day after her high school prom; i would like to have a few more detail how you did. i am a ham radio operator, most reliable contact is kb6jdy, 2839 fairgreen, arcadia, los angelus, ca, 91006
“Ubiquitous” is relative. 360s were “everywhere” but only in the limited spread of comouters at the time.
Don’t forget the XT/360 putti g a 360 sort of on the desktop via modified 68000s.
I’m actually working in a mainframe shop right now – we just migrated to a z15, and terms like DASD and IPL are still very relevant. They’re fantastically fast these days, but still show the heritage of punch cards and tape drives in everyday operation. There’s over 250 Linux virtual machines running alongside the CICS applications.
Yep. One of my clients has 2 z15s plus two new LinuxOne machines (which is basically z15 but linux IFLs only) and they couldn’t imagine life without
IBM System/360 referred to the circle (360 degrees) of applications it could be used for.
So don’t forget its derivative, the IBM System/4 Pi : Refers to 4 Pi steradians in a sphere: it was used for 3-D aero and astro navigation, made famous in both the space shuttle and F-15 flight computers, among many others.
neat. now do system/370 and 390 :)
Jim Nelson is correct. IBM mainframes are very relevant in large companies. We have 3 z15’s running tens of millions of transactions a day along with hundreds of z Linux guests. Z/OS can trace its roots back to the OS that ran on the 360’s. We support over 40,000 user on these 3 boxes.
I started on an IBM 360/50 in college and have worked on just about every model since. Including plug compatible models from Amdahl and Hitachi.
I too started on a 360 – and actually did have hearing protection in one shop – but people who think they (and the architecture) are dead and buried are very mistaken. I have a remote console up on another screen right now doing something with cics – and most of the stuff on this machine is very close to what I’ve been using for the last 45 years – indeed I have clists from the early 80’s that I still run very often…
When in college I was a p/t operations manager. For a 360/22. Yep, vacuum tapes, disk drives, printers that could play songs (with the right output). If I remember right, the dials had to be turned to 1-8-0 to Identify the disk drive to IPL from. In addition to the a/c, had to watch the humidity too.
When I hit college, the institution had upgraded from a 360 to a 370/VM/CMS which was the machine that was miniaturized when Intel came up with the 8080 chip…so I can’t say I’m surprised that the 360 has been degrading on mothballs. Still had to deal with big box hard drives, but also dealt with core memories, and one had to learn both hexacedimal and octal when it came to machine language because interfacing in the industrial world wasn’t standardized in that respect. Punch cards, magnetic tape on big reels, spooling routines in “Academic Computing” because peripherals were in different buildings…and yeah–the raised floors because of the controled environment similar to a cleanroom, and you had a smoorgasbord of printers to choose from: dot matrix, line, Selectric, plotters, card punchers, and forget ASCII–it’ was EBCDIC.
the mainframe is far from dead. IBM z16 can trace its lineage directly to the 360. programs from the ’60’s can still run on the latest mainframe. IBM has the majority of the mainframe market. In todays world two-thirds of the Fortune 100, 45 of the world’s top 50 banks, eight of the top 10 insurers, seven of the top 10 global retailers, and eight out of the top 10 telcos rely on mainframes for critical processes. how dead is that?
“THE IBM SYSTEM/360 REMEMBERED”
No thank you. I would prefer to forget it. And the system 36. And the AS400. 😉
As a CE in the 60s my first full system assigned to me was a S360/40. 1-2841 DASD Control Unit, 1-2311 Disk Drive, 2-2803-2 Tape Control Units, 16-2401 Mod 2 Tape Drives, several 2821 Control Units with 8-1403 Mod N1 Printers and 2-2540 Card Reader/Punches, 2-1012 Paper Tape Punches and 1-2671 Paper Tape Reader
As many have said the mainframe is by no means dead and has a few more of its lives still to go, as it has been pronounced dead a few times, and certainly if you want your credit card to work you better hope it does not die to soon. The line that does always come up in these articles is the raised floor and the large air-conditioning units. I am sure that we all know the server farms use huge amounts of power and generates masses of heat that needs A/C and has rows of racks sitting on top of raised flooring. That’s without bitcoin and the like.
There were buildings heated solely by their computing equipment.
It’s just called the “cloud” now. I do think that eventually people would stop trusting the cloud too much (after loads of cloud based stuff stops working) but there are a lot of things that just don’t decentralize easily.
Ah yes, the 70-hour weeks I put in as a tech writer/editor for the manuals we developed (using an amazingly sophisticated word processing program we were writing on the go and punched cards) to ship with the first 360s, followed by intensive courses in machine language and operation, Fortran, Cobol, APL, punch cards, core dumps, developing industry-specific expertise to automate our customers’ manual operations,…. I remember it well, and lovingly, despite the grueling hours and ever-increasing demands.
Someone from IBM said their only problem back then was a shortage of semi-log graph paper to use to plot their profits.
I worked on IBM mainframes for basically the first half of my career, and could write and debug code in BAL when needed. There were two principal operating systems, historically called DOS/VSE for smaller mainframes, and MVS for the larger ones — both were rebranded as “zServers” I believe in the 90’s. And then there was the VM/CMS server — VM was the first virtualized operating system — it started out as a way for DOS and MVS OS engineers to share one mainframe but work with their own copy of an operating system to engineer. VMware and the technologies that drive cloud computing is based on the VM model.
And I will claim to be the last system administrator in the USA who used punch cards as part of the computer boot-up .. this until 1986? Anyone more recent that can take this title? ;0)
Jim Vence, back in the late 1990s I was informed of a computer service company operating out of Phoenix AZ that was still doing punch card services for companies still using punch cards currently to that time, so somebody out there has you beat as the last card puncher. I was given to understand that most of those that required punch card services in the 1990s were governments that never got budgets for new equipment. Not long after that (1998? something like that) I visited a government computer installation at Sandia NM and they were still using humongous magnetic tape reels on mainframes. You guessed it–no funding for updating computer equipment there either.
The systems could share DASD snd tape drives, but there were limits on cable lengths. At one point our shop had two 65s, two 50s snd a 195. To allow sharing, the DASD and tapes were on the floors above snd below the floor with the CPUs. Of course there was also a floor of 026 and 029 keypunches.
Don’t forget the manuals, Principles of Operation, Messages and Codes. I once shut down s Moonie who eas trying to recruit me in the library when she asked me what I was reading and I snswered ” The IOS PLM. A freibd fought off s mugger with Messages and Codes.
The Principles of Operation, which was about 160 pages for the original System/360, is now about 2200 pages for the latest Z series machine.
Sorry, I know this is about 360s, but my father worked for IBM at the 1930 census in Baltimore. There was 1 Customer Engineer for every 6 keypunches. He was shocked when he visited my workplace in 1972 and saw me fixing a syntax error by holding the card I was duping to prevent it from advancing while I inserted s character.
I started at IBM in 1967 at the age of 19 in Production Control in Greenock, Scotland when the 029 and 059 card punch and verify machines were at their zenith, working with pen and paper for the most part, good times. I can still recall quite a few of those machines’ part numbers, trying to get my hands on some of those parts at the time presented challenges!!! And then came System 360 and lots of things changed. I retired from IBM Raleigh, NC in March 2014 after moving to the USA in 1998, the only job I ever had, 46 years and 5 months, those days are long gone sadly
I learned programming on a 360 system in the early 70s. I remember our first day in class the instructor told us we would be working with an IBM 360, and he stuck his chest out and continued “with 64K of core memory.” I flashed on that one day as I booted up my $200 Commodore 64.
I joined IBM right out of college in 1968. At that time, they were hiring anyone who could spell “programmer”. Indeed, we didn’t have a computer at the college, nor did they give any programming courses, but IBM offered me $150/week which was more than other offers I received having a BS in Biology.
I began my career at the ASDD Mohansic Lab in Yorktown, NY. I was immediately set to Basic Programmer training in Kingston, NY. There we were introduced to Assembler, PL/1, hexadecimal, EBCDIC, and JCL. We got our own S/360 Reference Data cards – the ubiquitous “green card”.
Back at Mohansic, I worked in a support group writing programs for various projects. We had a single S/360 Mod 40 with 256K of core memory. I learned how to load cards in the reader, mount tapes and disks, and flip switches on the operator console to debug code. It was great fun!
dec 1964 I was hired by IBM in Poughkeepsie. We were manufacturing the SLT boards used in the then System 360 system.. Many long hours but a great experience