This was posted 14hours ago on Instructables:
NOT, AND, OR gates using NAND gates - All
Any other recollections of our earliest computers?
Did any one work with analogue computers before the digital variety became common?
Anyone work with the sons/daughters of Collosus or Eniac?
Did anyone know Westjim?
NOT, AND, OR gates using NAND gates - All
Any other recollections of our earliest computers?
Did any one work with analogue computers before the digital variety became common?
Anyone work with the sons/daughters of Collosus or Eniac?
It was common back then to leverage DeMorgans theorem. The Apollo flight computers on the lunar lander only used one sort of gate as only one had been qualified for space.
Maurice Wilkes was my hero. Great that EDSAC has been rebuilt. Science museum as well as the computer museum at bletchley should be on your to see list if you ever get that far south.
Maurice Wilkes was my hero. Great that EDSAC has been rebuilt. Science museum as well as the computer museum at bletchley should be on your to see list if you ever get that far south.
The supercomputers designed by Seymour Cray (CDC-6600, Cray-I, Cray-II) all used NOR gates exclusively.
In the case of the 6600, this was because the {discrete!} DCTL logic family ran at such low signal swing (~1.3*VBE) that only NORs were possible. LINK
In the case of the Cray-I, this was because the {integrated circuit} ECL logic family used, gave significantly faster NOR gates than AND, NAND, AOI, or OAI gates.
It's worth mentioning that the first commercially available digital logic ICs, called Fairchild Micrologic, were NOR gates. uL914 was the part number. They superficially resemble the DCTL NOR gates used in supercomputers, but they ran at significantly lower power, significantly wider signal swings, had significantly bigger noise margins, and ran significantly slower. Not everybody is trying to build a liquid cooled supercomputer.
In the case of the 6600, this was because the {discrete!} DCTL logic family ran at such low signal swing (~1.3*VBE) that only NORs were possible. LINK
In the case of the Cray-I, this was because the {integrated circuit} ECL logic family used, gave significantly faster NOR gates than AND, NAND, AOI, or OAI gates.
It's worth mentioning that the first commercially available digital logic ICs, called Fairchild Micrologic, were NOR gates. uL914 was the part number. They superficially resemble the DCTL NOR gates used in supercomputers, but they ran at significantly lower power, significantly wider signal swings, had significantly bigger noise margins, and ran significantly slower. Not everybody is trying to build a liquid cooled supercomputer.
The first stored-programme computer (the Manchester baby) ran its first programme in 1948, so 1971 machines hardly qualify as "our earliest computers". I was born in 1970, so I have no recollections of early computers, I just read a book and a few articles about them.
If you look at Alan Turing's 1945 design on which the ACE series of computers was based, it actually uses logic cells that are rather similar to neurons with fixed and equal synaptic weights: the output of an n-cell becomes 1 when n or more of its inputs are 1. I believe the same holds for the earliest American designs. Some cells have an inhibit input, the output always remains 0 when inhibit is 1. This is not like AND or OR gates at all, except for the cells that have no inhibit input and have n equal to 1 (OR-gate) or equal to the number of normal inputs (AND-gate). A 0-cell with inhibit is an inverter.
If you look at Alan Turing's 1945 design on which the ACE series of computers was based, it actually uses logic cells that are rather similar to neurons with fixed and equal synaptic weights: the output of an n-cell becomes 1 when n or more of its inputs are 1. I believe the same holds for the earliest American designs. Some cells have an inhibit input, the output always remains 0 when inhibit is 1. This is not like AND or OR gates at all, except for the cells that have no inhibit input and have n equal to 1 (OR-gate) or equal to the number of normal inputs (AND-gate). A 0-cell with inhibit is an inverter.
Last edited:
I had Cray I calculation time-spots at A'dam U in the early '80s. (SARA)
Back then, one had to stand in line. Somewhat amusing, considering the calculation power/speed of current 500 buck pc's.
Vacuum tubes, with their enormous voltage swings, can build either NOR gates or NAND gates. An N-input NOR gate requires N triodes, while a J-input NAND gate requires only a single tube with (J+1) grid electrodes. However it was quickly discovered that ALL-tube logic was wasteful; far better to perform the logic function (OR, AND) with "crystal rectifiers" and then follow that with a single-triode level restoring inverter. Now you can build high fanin logic gates using only a single tube and N crystal rectifiers. The semiconductor logic family DTL copied this idea, 15 years later, when diodes were cheaper and easier to build than transistors.
Threshold logic (out=0 if and only if count(input 1's) >= K) didn't last long, because
Threshold logic (out=0 if and only if count(input 1's) >= K) didn't last long, because
- the noise margins of an N-input threshold logic gate are smaller than the noise margins of a binary logic gate; smaller by a factor of (N-1). Thus binary logic is more robust than threshold logic, by factors of many.
- boolean algebra is a simple and powerful design tool for binary logic gates; no powerful design tool is available for threshold logic
That's just one way of making logic gates with tubes/valves; because of its high clock rate (1 MHz), the 1950 pilot ACE used current steering logic and voltage followers, essentially an early type of ECL.
In started studying Engineering in 1969 at private Buenos Aires Catholic University and we used a donated IBM1620,
which was programmed with punched cards or tapes (I still keep a couple) and output was "bedsheet" sized fanfold paper ... no screens involved.
State run "Buenos Aires University" had a much larger and way more powerful IBM360, the typical "James Bond movie" evil computer:
, I suspect they didn't have it (physically) in the building, but it was remotely run by VT100 Terminals, while we could actually get inside the air conditioned room and touch the beast with our hands.
We wrote our own programs, of course: FORTRAN language for our Engineering needs, BASIC for mundane tasks (including ... gasp! ... silly games !!!!!!!!!!!)
We despised the Economics Faculty guys who used "petty shop attendant's" COBOL .
which was programmed with punched cards or tapes (I still keep a couple) and output was "bedsheet" sized fanfold paper ... no screens involved.
State run "Buenos Aires University" had a much larger and way more powerful IBM360, the typical "James Bond movie" evil computer:
, I suspect they didn't have it (physically) in the building, but it was remotely run by VT100 Terminals, while we could actually get inside the air conditioned room and touch the beast with our hands.
We wrote our own programs, of course: FORTRAN language for our Engineering needs, BASIC for mundane tasks (including ... gasp! ... silly games !!!!!!!!!!!)
We despised the Economics Faculty guys who used "petty shop attendant's" COBOL .
My little stash of E180CC double triode tubes comes from a digital control systems I found in the '80 at a scrapyard where mainframes came to die. A the time, environement protection was less a concern than today and this surreal place was basically a open air (muddy) field littered with the remains of military devices and partially dismantled IBM, DEC and Honeywell racks. A few years later, I read on the newspaper that the place was closed and the owner fined for the pollution.
This unlabeled control unit was next to the rusty remains of some mechanical pheriperals and had a early '60 look; it was a cluster of 6-7 full height racks. Part of the drawers of the racks contained giant rotary relays (the kind of relay used on electro-mechanical exchanges), interconnected with wire bundles. I assumed they where the sequential logic of a state-machine. Other drawers contained rows upon rows of E180cc tubes seemingly arranged as flip-flops (a work memory area, maybe?). On several drawers the tubes were arranged as decadic counters, each tube was near to a neon bulb on the control panel. Surely a complex machine, designed with minimal o no automation support: back then, designers had no VHDL compilers. Thousands of electrical contacts, this should have been a nightmare to service and repair. Anywhay, I'glad to have recovered a full bag of perfectly fine Philips E180CC for a very small amount of money.
This unlabeled control unit was next to the rusty remains of some mechanical pheriperals and had a early '60 look; it was a cluster of 6-7 full height racks. Part of the drawers of the racks contained giant rotary relays (the kind of relay used on electro-mechanical exchanges), interconnected with wire bundles. I assumed they where the sequential logic of a state-machine. Other drawers contained rows upon rows of E180cc tubes seemingly arranged as flip-flops (a work memory area, maybe?). On several drawers the tubes were arranged as decadic counters, each tube was near to a neon bulb on the control panel. Surely a complex machine, designed with minimal o no automation support: back then, designers had no VHDL compilers. Thousands of electrical contacts, this should have been a nightmare to service and repair. Anywhay, I'glad to have recovered a full bag of perfectly fine Philips E180CC for a very small amount of money.
Depends if you consider Tommy Flowers to have got their first 🙂
If you look at Alan Turing's 1945 design on which the ACE series of computers was based, [/QUOTE]
Do you have any references to this? There were various articles on how useful the pilot ACE was for analysing aerial photos and I've stared at the example in the science museum but not found anything to tell me how it worked or was programmed.
For some retro programming there is an EDSAC simulator http://www.dcs.warwick.ac.uk/~edsac/
My favourite Maurice Wilkes quote "The EDSAC was on the top floor of the building and the tape-punching and editing equipment one floor below. [...] It was on one of my journeys between the EDSAC room and the punching equipment that "hesitating at the angles of stairs" the realization came over me with full force that a good part of the remainder of my life was going to be spent in finding errors in my own programs. "
To put this in context people build the early computers without putting that much thought in to how to program them!
"early" computers. I watched IBM remove the vacuum tube powered 709 from UCLA's "Western Data Processing Center", and replace it with a 7094, no tubes. I used to have one of the "plugable units" with a magnetic core memory device. Long gone. My mom worked there, and helped write the PL/1 user manual. Early '60's maybe?
When I was in college I used a KIM-1 (6502 single-board computer, now collectible!) and the then-brand-new Apple ][, as well as dialup to a mainframe.
What was most "fascinating" was the Bitran Six, a large tabletop machine using (presumably) transistors and core memory. It looked "old" in 1977, it was possibly from the early 70s or even late 60s. It was a six-bit machine with maybe 100 words of memory, made specifically for teaching about computers. Programs were loaded and run with toggle switches, and (incandescent) lights were used to display register contents (the actual register flip-flops drove the lights).
What was most "fascinating" was the Bitran Six, a large tabletop machine using (presumably) transistors and core memory. It looked "old" in 1977, it was possibly from the early 70s or even late 60s. It was a six-bit machine with maybe 100 words of memory, made specifically for teaching about computers. Programs were loaded and run with toggle switches, and (incandescent) lights were used to display register contents (the actual register flip-flops drove the lights).
Younglings,
This is the first computer I got to pay with. It did have a high speed paper tape reader to load the operating system.
https://en.m.wikipedia.org/wiki/Bendix_G-15
Next up was a PDP8S. Eventually bought a PDP8I for my first computer.
ES
This is the first computer I got to pay with. It did have a high speed paper tape reader to load the operating system.
https://en.m.wikipedia.org/wiki/Bendix_G-15
Next up was a PDP8S. Eventually bought a PDP8I for my first computer.
ES
I've also stared at it quite a bit; I was on vacation in England this year and finally had a chance to see the pilot ACE, the rebuilt Colossus and a rebuilt Turing-Welchman bombe.
Anyway, Colossus definitely predates the Manchester baby, but with a stored-programme computer I mean a computer that is designed to have its programme stored in a similar type of memory as the data, as opposed to computers that are meant to be programmed by flipping switches and rewiring stuff (even if by rewiring them in a smart fashion, you can make them run a programme from memory).
There is a fascinating book about the ACE series of computers edited by Jack Copeland:
B. Jack Copeland (Editor), Alan Turing's Automatic Computing Engine – The master codebreaker's struggle to build the modern computer, Oxford University Press, 2005, ISBN 0 19 856593 3
A lot of information can also be found here:
AlanTuring.net
AlanTuring.net A Brief History of Computing
Alan Turing's system design for the ACE can be found here:
AlanTuring.net
It was considerably simplified by Turing and others and the electronic circuits were completely redesigned before it became the pilot ACE. By that time Turing had already left the NPL and moved to Manchester, to work on the Manchester computer. The ACE project was delayed enormously by unavailability of most of Tommy Flowers's team (too busy restoring the British telephone network after WWII), by problems getting the mercury delay line memories to work reliably and by plain mismanagement, otherwise the ACE would probably have been the first stored-programme computer.
Last edited:
Below is an interesting link about early "minis".
I find the video showing the actual operation of the German LGP-30, a drum-storage machine, to be pretty amazing!
Who Built the First Minicomputers?
I find the video showing the actual operation of the German LGP-30, a drum-storage machine, to be pretty amazing!
Who Built the First Minicomputers?
That's the first pic I've seen of a Scientific Data Systems computer. That company was discussed in the book "Dealers of Lightning" (covers mostly the 70s to early 80s, about the happenings in Xerox Parc - of the computer history books I've read, if you're interested in where the Mac and MS Windows came from, I recommend this one most highly). Xerox opened Xerox PARC, and they needed computers, but rather than ask those doing the research what they needed, they bought SDS. Those at PARC thought it stank but since Xerox owned it they couldn't get a PO to buy any other computer (they wanted a DEC), so they built a DEC knockoff. They named it the Alto, added a bitmapped display, and the rest is history, as shown here:
https://www.youtube.com/watch?v=RFKCZLwbwwg
The last words are prophetic: "...[Xerox] could have been the IBM of the 90s, could have been the Microsoft of the 90s."
When I was a kid in the 1950s, we subscribed to some series of "Science Kits", and one of them was an "Analog Computer". It wasn't very precise, it was made in a cardboard box with a printed cardstock panel, and several dials with legends. It had a battery, a pot with pointer knob in the center of each dial, and a little current meter. In retrospect, it was essentially a slide rule. You could set two numbers on two dials, then rotate the third dial until the meter nulled, then read out the result on that dial scale.
- Status
- Not open for further replies.