Contrary to popular belief, yes you can. In fact, many existed and did very important. Perhaps one of the most well known is the ENIAC (Mauchly pronouned it En-ee-ack not EEn-ee-ack who can argue with its primary designer?).
ENIAC
The ENIAC had flip-flops stored in a ring counter that accumulated (or counted) pulses sent to it. Each ring had ten flip-flops that represented a value from zero (0) to nine (9) and there were ten rings in order to store ten digits (range is 0 to 9,999,999,999). One last flip-flop stored the sign. Each of the cabinets above that the ladies are busy programming (yes they are) is essentially analogous to a CPU accumulator which was an early type of register that always added the number it stored to anything “moved” or “stored” into it. Each cabinet has a complete circuit of 10 ring counters (later 20 digits). The ENIAC was a digital computer but not a binary computer as it used base ten.
Programming the ENIAC involed using a portable function table (you can see on to the right belwo that consisted of 1,200 switches to store numbers). Each cabinet (or multiple) were wired to perform some mathematical function and then the cabinets (now a particular math function or part of one) were wired together to perform the logic. It could branch and loop and yes it was hard to program taking several weeks for a single problem.
It had no type of memory at all. Just a bunch of what were essentially accumulating registers. Here’s what a ENIAC program looked like on paper:
What kind of work did it do? Notably ballistic wiring tables for artillery for WWII gunners. Oh yeah, and that little bit of work simulating how neutrons would likely travel through various materials to develop the first nuclear bombs.
And it was programmer mostly by a set of very smart gals. But if you go back and find reference pictures at the time anything “released” to government or public (later) had men. Imagine that? Male models standing in for women to give it “credibility”. Sheesssh! At least know we know the truth.
The ABC
If you start asking what was the first digital computer, you’ll start what old timers call “a flame war”. See Iowa University will say they did it with the Atanasoff-Berry computer — the ABC. The subject of “programability” (it wasn’t) and “Turing complete” (it wasn’t) will come up and someone will throw in that it was a hybrid mechanical computer (rotating drums of capacitors) and not a fully electronically siwtched digital one. It was completed (fully tested) in 1942 even before the ENIAC was started in 1943. The ABC had two important differences: 1) it used binary arithmetic and 2) electronic switching. U of I says they did it, many others (the majority) say the Quakers at UPenn did it with the ENIAC. Perhaps the Quakers were correct as thirty years later (1973) the ENIAC patents were invalidated (Honeywell, Inc. v. Sperry Rand Corp) because it was concluded by a US district court that the ENAIC was derived from the ABC. Winner in court: UPenn. However the ABC was not an electronically switched programmable digital computer and the ENAIC was not a binary digital computer, both them operated without RAM.
The ABC had two drums of 1,600 capacitors that rotated once per second. They were organized into 30 units that could perform an addition or subtraction. It represented numbers in a 50-bit fixed-point numbers and could only be used to solve linear equations. This was known as regenerative capacitor memory which was distinctly not random access. But it did have the ability to stored 3,000 bits of information which you might consider memory but not RAM.
The Colossus
So while the Quakers and the Hawkeyes were flaming over who invented the first electronic digital computer the Britts on the other side of the pond released their top secret information about a system called the Colossus used to break cryptographic codes in WWII. It was designed by a postal worker (I’m not kidding!) and completed in 1943. It was programmable and fully electronic except for the program input which read “instructions” from a paper tape (mechanical). Even so all can agree it was the first programmable, electronically switched, digital computer. (Upenn retain their sly grin of “we were binary”). It was not a general purpose computer (Quakers sly grin here) since was designed to solve cryptographic problems surrounding the Lorenz SZ used by the German army in WWII, it was programmable using the K2 board. Of course, I could go further back with the Bombe but I think we can conclude it is not a digital computer fairly easily.
Okay, so we’ve looked at a bunch of stuff form the early days that many will say, “Those aren’t even computes.” Well, you’re wrong. They were and they performed many useful functions without using RAM. Let’s move into the days of memory.
The World of RAM
Williams Tube Memory
Here a CRT was used to store the bits. Like any CRT an electron beam blasted the tube and created a grid of dots that caused a build up of static electricity. A metal plate on the end of the tube was used to read the memory by detecting the sight charge of static electricity. Each charged lasted about 1/5 of a second so the CRT had to be constantly refreshed. But this was the first random access memory even though the bit density was small. Since the CRT used the phosphor as normal, each bit was visible except that you couldn’t see it because of the metal plate used for reading it.
(two tubes in the above unit)
Different types of memory came along including the astoundingly interesting mercury delay line that sent a stream of bits as a sound wave through a tube of mercury. The sound transmitted as a wave through the mercury. It had to traverse the full length of the tube before it could be read so it was serial access and had a delay. But these tubes might hold 512 to 768 bits of data giving those early computers 256 to 512 bytes of RAM - non-random access of course. Data literally went in one end and came out the other. See Mercury delay line memory. Once you understand it you’ll realize it was bits moving through the liquid metal method in a stream that were read one after the other. Moving bits of shiny metal.
Core Memory
This was all the rage when it came out. It used little ferrous rings and lots of copper wire to store a single bit in each ring which was magnetized either clockwise or counter clockwise. It used two sets do drive lines, a sense line, and an inhibit line to store a bit in each ring. It used a destructive read that flipped a set bit to zero which meant it had to be reset after a read (only for bits (1) not unset bits (0)). This was RAM because any single bit could be read in the system in a random fashion. This were at first fairly large (this memory module held 8 bytes):
but became fairly small (this held :
It was cheap and reasonably fast RAM and so the whole idea of a RAM being an essential component of a computer became universal. Now it was all about RAM.
The 6508
The MOS 6508 was a version of the 6502 that had 256 bytes of memory built into the the CPU that mapped to pages $00 and $01 for the zero-page and the stack. The stack grew downwards and needed maybe 64–96 bytes for a stack with a complicated programs (8 levels deep of subroutines). Essentially you could almost think of these as slower registers. It didn’t use RAM but it did have an 8-bit input/output port mapped to location $0001 with the direction at $0000. One single page of 256-bytes was all you got. But could be a whole hand-held calculator, a floppy drive controller, etc.
It is not unique there were other chips with memory onboard but the 6508 is the one I am most familiar with.
Oh and what about the world’s first truly fully electronic programmable binary digital computer? It wasn’t any of those above. It was the Z3. Sorry Quakers! Sorry Hawkeyes! Sorry Bletchley Park! Zuse beat you all — you’re squabbling over seconds. The Z3 world’s first electronically switched digital binary computer that was shown in 1997 to be Turing complete (although somewhat in a round about way). It was destroyed by bombing before the end of the war. Well as they say, the victors get to write history.
Comments
Post a Comment