|
In 1956 magnetic core memory was patented. Anyone else around here remember working on systems with "donuts" for memory?
>64
Some days the dragon wins. Suck it up.
|
|
|
|
|
theoldfool wrote: Anyone else around here remember working on systems with "donuts" for memory?
On this dinosaur pasture? Of course!  I bet you'll even find people who worked on drum memories.
My first encounter with core memories dates back to high-school when an intrepid teacher took us to see a factory of core memories. I remember being amazed at the speed and precision of those ladies (they were all or mostly ladies working there) who weaved those memory frames.
Mircea
|
|
|
|
|
Mircea Neacsu wrote: you'll even find people who worked on drum memories.
Yes, I did. But the first system I worked on had only mag tape, no rotating storage. And then we had the 1004 which used a backplane plugboard for its programs, and 1024 bytes of memory.
|
|
|
|
|
Our University had a 'museum corner' with GIER computer. One notable feature was the wooden teak side panels, another that one side wall was an array of large ferrite cores, with a mesh of lines either going through or bypassing the cores. That was the microcode of the machine: You could rewire the side wall to give the computer a modified instruction set!
If my memory is right, the GIER had 1024 words of 42 bits - an integer 32 bit part plus a 10 bit scaling factor. It also had a complete ALGOL60 compiler, split into 14 passes, using the drum for to transfer intermediate results from one pass to the next. The function of the last pass was to reorganize the blocks on the drum for more efficient execution of the program.
This machine was operational at the time I started my studies, but it was considered 'retired'; it was run only as part of 'history tours'.
|
|
|
|
|
From what year was that machine?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
I had to look that up: Norwegian Wikipedia[^] has a page about the GIER, telling that the machine to NTH (now named NTNU), the Norwegian Technical University, was delivered in 1962. I saw it 16 years later as retired; I do not know when it was retired.
The Wikipedia article also tells that it had no interrupt system at all. Its drum had 320 "channels" (tracks), each of 40 words of 42 bits. It had sort of a "software virtual memory" - system provided functions for sending blocks from the core primary memory to the drum to make room for another one retrieved from the drum. The paper tape reader was optical, reading 3000 chars/sec (!). Linpack speed: 0.66 Kflops  Weight: 2600 kg.
56 machines were built, and the machine is named from the first customer: Geodætisk Instituts Elektroniske Regnemaskine (Geodetic Institute's Electronic Calculator)
"Fun fact": The memory was built with 42 core boards, each of 1024 bits - each board providing one bit to the word. I guess that made it rather difficult to extend the architecture to a larger (or restrict it to a smaller) memory size.
An open source software GIER simulator is available for Linux/X11 and Windows. Expect most information to be in Danish. Assembler mnemonics was based on Danish terms, and in the Wikipedia assembler code sample, comments are in Danish as well. The Algol compiler was modified to accept æ and ø in identifiers.
|
|
|
|
|
trønderen wrote: another that one side wall was an array of large ferrite cores, with a mesh of lines either going through or bypassing the cores.
That was called a Core rope memory[^] most famously used on the Apollo Guidance Computer. In the case of AGC, the advantage of core rope memory was its immunity to cosmic radiation.
Mircea
|
|
|
|
|
A PDP-10 was our main system when I was at university. It had magnetic core memory, and I seem to recall that it used a drum for swap space.
|
|
|
|
|
Yup. I keep wondering if those were the "good old days". Worked on an old accounting machine, 7 words of memory, BCD, and people actually did payroll with it. Typewriter or punch card output, electronics were vacuum tubes (I have one on my desk).
Much later, did VAX at the University via 300 baud dial up.
>64
Some days the dragon wins. Suck it up.
|
|
|
|
|
I spent 1975-76 at a US high school, learning my first BASIC. But the school couldn't afford a "high speed" modem - that is a 300 bps one - so we had to make due with a 110 bps one. That also implied that our terminal was an extremely noisy, military style 10 cps Teletype, located in a huge all-concrete basement bomb shelter with no sound dampening whatsoever. (The high speed 300 bps modems usually came with a a more modern, lightweight and quieter terminal.)
You bring back memories of the 'baud' term! I belong to that generation considering 'baud' a synonym for 'bits per second', long after we got (truly) high speed modems where this certainly does not hold.
|
|
|
|
|
theoldfool wrote: I keep wondering if those were the "good old days".
No, the good days are now. We are just nostalgic for the days when we were young, strong and enthusiastic.
Back in the day I had incredible eyesight and monitors had 160 by 100 resolution. Now I need glasses for almost anything and my laptop comes standard with 1920 by 1024 resolution. Sometimes it feels like the world is thumbing the nose at me
Mircea
|
|
|
|
|
In the days when people still remembered drums, I picked up a Hacker's Dictionary explaining 'Fastrand' (a much used drum device from Sperry-Rand) as 'A device for storing angular momentum'.
When I started my computer studies, the University had just retired its Fastrand drum, but it was still sitting in the machine hall (along with two huge Univac 1100 series mainframes), so we got a chance to see it. The service engineers told us that it would be recycled as a steam roller
|
|
|
|
|
Colour me a drummer!
Looked like the artificial brain in a Pertwee-Era Dr Who AFAIR.[^]
So old that I did my first coding in octal via switches on a DEC PDP 8
modified 1-Mar-22 3:44am.
|
|
|
|
|
My high school had an HP system that had 4K of donut RAM. Also had a display (single line of LED characters, 40 max), optical card reader, thermal paper printer and a plotter.
|
|
|
|
|
I've got two blocks of core sitting on my desktop (the physical one), the biggest one holding 1152 bits. Junior programmers are fascinated when I invite them to inspect my core dump
In a 1978 summer job, the last summer before I started my studies, we were working on a mix of core and semiconductor RAM based machines. The group leader insisted that semiconductor RAM would be a short-lived fad - machines that loose their memory contents when power is turned off will never be satisfactory for the typical computer user. Software for these machines were distributed on 8 track punched tape from paper, aluminized mylar (keep your fingertips away from its edges when it runs!) or a fiber-enforced plastic material resembling those paper-looking but super-strong envelopes - I do not recall what that material was called, but no man could tear it, you needed a cutter / scissors.
|
|
|
|
|
I was late to the party but the place I worked had memory core boards for DEC systems as spares. I've held one in my hand and marveled at the struggles the ancients must have had when we advanced race of nerds had TI4464 64k x 4 chips at our disposal. And used those high and mighty 74ls181 ALUs to crunch what was stored there.
Now even that "sophisticated" stuff is a riot. 
I like to take walks down memory lane on ebay. They still want more that they should for some of it.
|
|
|
|
|
In college I took a class that used PDP-11/05's running RT-11. The bootstrap code was stored in about 80 words out of a 128 word core memory. Student code routinely ran wild and wrote over the bootstrap code. The bootstrap could be re-entered using the front panel switches on the machine. I had to do it once  .
There was one guy who had to do it so often, he could re-enter the 80-word bootstrap in just under 60 seconds. 
Software Zen: delete this;
|
|
|
|
|
Gary Wheeler wrote: Student code routinely ran wild and wrote over the bootstrap code.
One college text book, in a discussion of protection mechanism in hardware memory management systems told of of a machine with limited protection. Forgetting to initialize indexing variables, one student inadvertently sorted the machine monitor.
One of my fellow students dryly remarked: They had the same tings done, just in a somewhat different order, didn't they?
|
|
|
|
|
I always remember someone who brings donuts.
|
|
|
|
|
Yes, in fact in the computer junk store in Palo Alto, when I was 20 or so (39 years ago) I found a large magnetic core "board". Bought it but never did anything with it. Not sure where it disappeared to over the years.
|
|
|
|
|
Yeah, and the code for the system that used it was on paper tape.
"Time flies like an arrow. Fruit flies like a banana."
|
|
|
|
|
IBM System/3 Model 6 in 1972. I peaked inside and saw the donuts. I was merely the operator/key puncher while in school. It was programmed by another in RPG. - Cheerio
|
|
|
|
|
I do.
My first computer was an IBM 1620.
|
|
|
|
|
I’m a 60 year old coder. I’ve been doing it since I was 20. I always knew I loved it, but I just realized I truly do it for the get-up-and-walk-around moment. That’s the moment when you fix a sneaky bug, or complete a demanding or tricky task/algorithm/approach. It’s so satisfying, that you can’t simply move on to the next thing. You have to get up and walk around to bask in the satisfaction.
I’m chasing a get up and walk around worthy bug in a bit of embedded C++ at the moment. It’s a timer fringe case, or a variable the should be volatile, and it’s not. I’ll get it, and I’ll certainly need to get-up-and-walk-around once it’s dead.
Has anyone else recognized the need to get-up-and-walk-around after a truly satisfying coding moment? Do you have other victory rituals?
|
|
|
|
|
I hadn't noticed it but yeah I do get up and walk around after a big sign of relief.
The thing about C/C++ is you can shoot yourself in the foot and the problem is so subtle that it takes special skills and a lot of cursing to find and fix.
The less you need, the more you have.
Even a blind squirrel gets a nut...occasionally.
JaxCoder.com
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
