a186@mindlink.UUCP (Harvey Taylor) (06/07/90)
In <3040@softway.oz>, adjg@softway.oz (Andrew Gollan) writes: > >I have always liked the idea of mercury delay lines, and the story >someone told me of "playing them like a harp" when his program went >wrong and getting weird results from the machine. > >Can anyone tell me what machines had these memories? I was just reading A Few Good Men from Univac [ISBN 0-262-12120-4] last night & Lundstrom mentions that the Univac_I used Hg Delay lines. UNIVAC_II moved up to core memory. <-Harvey PS. So what did become of FRAM? "Man is a rational animal --- so at least I have been told." -Bertrand Russell Harvey Taylor Meta Media Productions uunet!van-bc!rsoft!mindlink!Harvey_Taylor a186@mindlink.UUCP
adjg@softway.oz (Andrew Gollan) (06/07/90)
I have always liked the idea of mercury delay lines, and the story
someone told me of "playing them like a harp" when his program went
wrong and getting weird results from the machine.
Can anyone tell me what machines had these memories?
--
Softway: +61 2 698 2322 Andrew Gollan adjg@softway.oz
GPO Box 305, Strawberry {uunet,mcvax}!softway.oz!adjg
Hills NSW 2012 Fax: +61 2 699 9174johnl@esegue.segue.boston.ma.us (John R. Levine) (06/07/90)
In article <3040@softway.oz> adjg@softway.oz (Andrew Gollan) writes: >I have always liked the idea of mercury delay lines, ... >Can anyone tell me what machines had these memories? They were popular in the 1950s. Some of the early British machines used them. The only commercial delay line machine I know of is the Packard Bell 250, later the Raytheon 250, which I have actually programmed. It was from about 1960, was the size of a large suitcase, and only drew 100 watts, less than the console Flexowriter. It had almost no console lights, and you did everything from the typewriter using an octal debugger. (Well, if you wanted to, you could put a scope on a line or register and adjust the triggering to see the word you wanted.) The memory was divided into some number of lines (about six) each of a few hundred 22 bit words. You could execute instructions in successive locations, but that required a full cycle of the line per instruction for an instruction time measured in milliseconds. It was much faster to spread your instructions along the lines to minimize the wait. I did most programming on big sheets of paper spread out on the floor. Line zero was 1/4 the normal length, so each word had four addresses which you could use to minimize the delay. The most peculiar think about it was the I/O addressing -- the line you executed the I/O instruction from determined what device you used, e.g. a write from line 5 went to the typewriter, and a write from line 6 went to the paper tape punch. And you thought Packard Bell just got into the computer business last year! -- John R. Levine, Segue Software, POB 349, Cambridge MA 02238, +1 617 864 9650 johnl@esegue.segue.boston.ma.us, {ima|lotus|spdcc}!esegue!johnl Marlon Brando and Doris Day were born on the same day.
kend@mrloog.WR.TEK.COM (Ken Dickey) (06/07/90)
In article <3040@softway.oz> adjg@softway.oz (Andrew Gollan) writes: >I have always liked the idea of mercury delay lines, ... >Can anyone tell me what machines had these memories? Maurice Wilkes's EDSAC, the first stored program to operate, used mercury delay lines [as reported by Arthur Burks in "A History of Computing in the Twentieth Century", Metropolis, Howlett, & Rota, Eds., Academic Press (1980)--ISBN 0-12-491650-3]. EDVAC was designed to use mercury delay lines and the design influenced a number of later computer designs. I don't know if it was ever built. Anyone? -Ken Dickey kend@mrloog.WR.TEK.COM
johnl@esegue.segue.boston.ma.us (John R. Levine) (06/08/90)
In article <2694@wrgate.WR.TEK.COM> kend@mrloog.WR.TEK.COM (Ken Dickey) writes: >Maurice Wilkes's EDSAC, the first stored program to operate, used mercury >delay lines ... I don't know if it was ever built. Anyone? Accoring to Wilkes, the EDSAC ran its first program on 6 May 1949, about 2 years after construction started, and worked until 11 July 1958. Wilkes' book "Memoirs of a Computer Pioneer," published by MIT Press is well worth reading for its insights into the origins of modern computing. Many people are surprised to learn that Wilkes invented microprogramming in 1951, though he didn't build a computer based on it until 1958. -- John R. Levine, Segue Software, POB 349, Cambridge MA 02238, +1 617 864 9650 johnl@esegue.segue.boston.ma.us, {ima|lotus|spdcc}!esegue!johnl Marlon Brando and Doris Day were born on the same day.
graham@convex.com (Marv Graham) (06/08/90)
EDVAC was built - with mercury delay lines - and ran for several
years the the Ballistic Research Lab an Abeerdeen Proving Ground.
Marv Graham; Convex Computer Corp. {uunet,sun,uiucdcs,allegra}!convex!grahamkend@mrloog.WR.TEK.COM (Ken Dickey) (06/08/90)
In article <1990Jun7.210822.5230@esegue.segue.boston.ma.us> johnl@esegue.segue.boston.ma.us (John R. Levine) writes: >In article <2694@wrgate.WR.TEK.COM> kend@mrloog.WR.TEK.COM (Ken Dickey) writes: >>Maurice Wilkes's EDSAC, the first stored program to operate, used mercury >>delay lines ... I don't know if it was ever built. Anyone? >Accoring to Wilkes, the EDSAC ran its first program on 6 May 1949, about ... Note that the `...' elides the reference I made to EDVAC [*not* EDSAC]. ^ ^ | | The EDVAC design was done by a group at the Moore School of Engineering along with von Neumann. It would not have made sense for the `first stored program to operate' never to have been built. Nits, nits, nits! -Ken Dickey kend@mrloog.wr.tek.com
chrisp@mks.com (Chris Phillips) (06/09/90)
In article <2694@wrgate.WR.TEK.COM> kend@mrloog.WR.TEK.COM (Ken Dickey) writes: >In article <3040@softway.oz> adjg@softway.oz (Andrew Gollan) writes: >>I have always liked the idea of mercury delay lines, >... >>Can anyone tell me what machines had these memories? > > >Maurice Wilkes's EDSAC, the first stored program to operate, used mercury > >-Ken Dickey kend@mrloog.WR.TEK.COM I remember going to a lecture ('80 Forsythe Lectures?) at Stanford by Wilkinson where he reminisced about programing this machine. I was working on microcode for a machine with a pipeline architecture at the time and found the techniques he described very similar to what we were doing, even to the sort of timing drawings they used to design their algorithms ! They had used the pipes to store partial results for computations and this allowed them to get very effective computational speed with SLOW hardware. Chris Phillips chrisp@mks.com -or- chrisp@telly.on.ca
nather@ut-emx.UUCP (Ed Nather) (06/09/90)
In article <1990Jun7.154830.3294@esegue.segue.boston.ma.us>, johnl@esegue.segue.boston.ma.us (John R. Levine) writes: > In article <3040@softway.oz> adjg@softway.oz (Andrew Gollan) writes: > >I have always liked the idea of mercury delay lines, ... > > >Can anyone tell me what machines had these memories? > > They were popular in the 1950s. Some of the early British machines used them. > The only commercial delay line machine I know of is the Packard Bell 250, > That machine used magnetostrictive delay lines, not mercury, which is why it was so small. The only machine I ever saw that used mercury delay lines for its memory was the original Univac; they were huge tanks suspended from the ceiling of the chassis (you walked into the chassis through a door). The tanks had vacuum tubes festooned all over them. I asked the tech how they ever kept it all running, and he repleid it wasn't very hard: if it failed, just walked in and looked for the tubes whose filaments were out (which you could see through the glass bulbs). Apparently the flipflops were so well designed, with such wide margins, they never failed unless the filament died and electrons quit flowing through the tubes. -- Ed Nather Astronomy Dept, U of Texas @ Austin
a186@mindlink.UUCP (Harvey Taylor) (06/11/90)
Speaking of old technologies, in A Few Good Men from Univac,
Lundstrom says(Page 22):
"Magnetic amplifier circuits were thought for a brief time
to be the most suitable replacement for vacuum tubes in computers.
They were superseded by transistors before much design was done
with them."
I have never heard of this technology before. Anybody got the facts?
<-Harvey
"History repeats itself; historians repeat each other."
Harvey Taylor Meta Media Productions
uunet!van-bc!rsoft!mindlink!Harvey_Taylor
a186@mindlink.UUCPmshute@cs.man.ac.uk (Malcolm Shute) (06/11/90)
>In article <1990Jun7.210822.5230@esegue.segue.boston.ma.us> johnl@esegue.segue.boston.ma.us (John R. Levine) writes: [about the Cambridge University machine] >>Accoring to Wilkes, the EDSAC ran its first program on 6 May 1949... In article <2701@wrgate.WR.TEK.COM> kend@mrloog.WR.TEK.COM (Ken Dickey) writes: [about the Moore School machine] >The EDVAC design was done by a group at the Moore School of Engineering >along with von Neumann. It would not have made sense for the `first >stored program to operate' never to have been built. True... but EDVAC wasn't the first either!!! The Manchester University Mark I machine was the first stored program machine to run. It first executed a program correctly (to compute greatest common factors) on 21st June 1948 (I believe just a matter of a couple of days before von Neumann's team got their machine working). >Nits, nits, nits! >-Ken Dickey kend@mrloog.wr.tek.com From me too. -- Malcolm SHUTE. (The AM Mollusc: v_@_ ) Disclaimer: all
aw1r+@andrew.cmu.edu (Alfred Benjamin Woodard) (06/11/90)
I've been following this thread for quite a long time and being a computer histroy / trivia buff I am curious what Mecury Delay lines were. It seems really weird reading about all the computers that used them and not knowing what they were, -ben
faiman@m.cs.uiuc.edu (06/12/90)
The mercury delay line was an early form of serial memory. Aside from EDSAC, another early British machine to use mercury lines was LEO (Lyons Electronic Office). Nickel magnetostrictive delay lines soon replaced the mercury type, as being smaller, cheaper, and less dangerous. I remember a machine at Elliott's back in the late 50's called Nicholas, on account of its nickel delay line memory. The Elliott 802 and 803, both serial machines, also used nickel lines, but only to pad out the lengths of their CPU registers; their memories were core. -- Mike Faiman, Urbana faiman@cs.uiuc.edu
sxr@cs.purdue.EDU (Saul Rosen) (06/12/90)
In article <EaQvAbO00Ug7A7a68t@andrew.cmu.edu> aw1r+@andrew.cmu.edu (Alfred Benjamin Woodard) writes: >I've been following this thread for quite a long time and being a >computer histroy / trivia buff I am curious what Mecury Delay lines >were. It seems really weird reading about all the computers that used >them and not knowing what they were, > >-ben The Mercury Delay Line is not of real current interest in the world of computer architecture, and there is not much point to including a technical description here. It is of course of great historical interest, and is discussed in various books on the history of computers. See for example "The Computer from Pascal to von Neumann" by Herman H. Goldstine, and "A History of Computer Technology" by M. R. Williams. The Mercury Delay Line memory was invented by J. Presper Eckert at the Moore School of Electrical Engineering of the Univ. of Pennsylvania in 1943-44. The concept of the modern stored program computer was introduced in the original design of the EDVAC in 1945. The serial design of the EDVAC was based on the characteristics of the Mercury Delay Line memory. Maurice Wilkes attended a special course on computers at the Moore School in the summer of 1946, and he went back to Cambridge in England and built a small Mercury Delay Line computer, the EDSAC. Eckert and John Mauchly left the Moore School and formed a company which became the Eckert-Mauchly Computer Corporation(EMCC). They built two Mercury Delay Line computers. The first was a relatively small one called the Binac. Both the EDSAC and the Binac were more or less finished in the spring of 1949. There is some argument about which was first. The EDVAC, much modified from its original design, was finished somewhat later. The most important Mercury Delay Line computer was the Univac, later known as the Univac I. It was built by EMCC which was absorbed by Remington-Rand. There were a total of 46 Univac I systems delivered. The Uni in Unisys is still a reference to Univac. For a number of years the Mercury Delay Line was the only really reliable computer memory available, and there were a fairly large number of computers built that used it. In the US there were the Raytheon Raydac and the SEAC built by the National Bureau of Standards. A number of universities built computers based on the SEAC, of which the best known was the Midac at the U. of Michigan. In England the ACE and Pilot ACE led to a commercial DEUCE computer using Mercury Delay Line memory that was marketed by English Electric.
yarvin-norman@CS.Yale.EDU (Norman Yarvin) (06/12/90)
In article <EaQvAbO00Ug7A7a68t@andrew.cmu.edu> aw1r+@andrew.cmu.edu (Alfred Benjamin Woodard) writes: >I've been following this thread for quite a long time and being a >computer history / trivia buff I am curious what Mercury Delay lines >were. It seems really weird reading about all the computers that used >them and not knowing what they were, A tube of mercury, with a speaker at one end and a microphone at the other. (actually, transducers at both ends.) Sound takes time to travel, so there is a delay between driving the speaker and getting a response. I presume mercury was chosen for its acoustic properties (high density for high sound energy, liquid for immunity from shear waves). -- Norman Yarvin yarvin-norman@cs.yale.edu
weaver@weitek.WEITEK.COM (06/12/90)
The magnetic amplifiers mentioned worked as follows: a strong clock pulse and a synchronous weak data pulse are both fed into a magnetic core. If the sum of the currents input are great enough, and opposing the current magnetic state of the core, the core reverses magnetic polarity, and a magnetic pulse is generated. A seperate output winding can be used to pick up this output pulse. The output pulse can be more powerful than the the data pulse, so the circuit is an amplifier. An Wang, who founded Wang labs did a lot of work on this area (in the 40s (?)) before magnetic cores were found to be useful for bulk memory, making logic gates and such. The only commercial use I have heard of was in one of the first core memories sold by IBM. Large cores were used in a decoder circuit to drive the memory plane. This reduced the tube count from (say) 32 to (say) 4. I learned this John Pugh's "Memories that Shaped an Industry", a book on core memories from by a former IBM engineer.
news@ism780c.isc.com (News system) (06/13/90)
In article <2072@mindlink.UUCP> a186@mindlink.UUCP (Harvey Taylor) writes: > > Speaking of old technologies, in A Few Good Men from Univac, > Lundstrom says(Page 22): > "Magnetic amplifier circuits were thought for a brief time > to be the most suitable replacement for vacuum tubes in computers. > They were superseded by transistors before much design was done > with them." In the late 50's, Sperry Rand (or may be it was still Remington Rand) manufactured machines called the "Solid State 80" and "Solid State 90". These machine were based an magnetic amplifiers. The only difference between the machines was that the 80 model processed standard 80 character Holerith code and the 90 model processed 90 charcter cards. The machine used a magnetic drum for its main memory. Marv Rubinstein
haynes@ucscc.UCSC.EDU (99700000) (06/13/90)
Magnetic amplifiers were used in at least one commercial computer,
the Univac Solid-State 80/90. (80 for 80-column IBM style punched
cards, 90 for 90-column Remington-Rand style punched cards). It
wasn't entirely solid state, as there was about a 2-kilowatt
vacuum tube clock supply to make it all go.
There was a paper in IEEE Transactions on Computers (or probably it
was still IRE back then) about the logic. Seems to me Univac has
had a history of taking some rather oddball kinds of logic and
using them in production computers.
At the time we didn't have transistors fast enough to be attractive
in real computers, so maybe that's why the magnetic amplifiers looked
temporarily promising.
haynes@ucscc.ucsc.edu
haynes@ucscc.bitnet
..ucbvax!ucscc!haynes
"Any clod can have the facts, but having opinions is an Art."
Charles McCabe, San Francisco Chroniclemike@mgh-znmr.harvard.EDU (Mike Vevea) (06/13/90)
Refering to magnetic amplifiers, In article <25079@weitek.WEITEK.COM> weaver@weitek.UUCP (Michael Weaver) writes: >The only commercial use I have heard of was in one of the first >core memories sold by IBM. I don't remember much about how they worked, but the Univac/Remington Rand SS-80 and SS-90, built in the late 1950s used magnetic amplifiers instead of tubes or transisters as the active element in their gates. There are a few others around (Ed Gould?? Mike Albaugh?? Can you correct my fading memories?). The computer club at UC Berkeley had a couple of these in the late 1960s/early 1970s. mikeV (mike@mgh-znmr.harvard.edu)
martin@minster.york.ac.uk (06/13/90)
In article <1317@m1.cs.man.ac.uk> mshute@cs.man.ac.uk (Malcolm Shute) writes: >>In article <1990Jun7.210822.5230@esegue.segue.boston.ma.us> johnl@esegue.segue.boston.ma.us (John R. Levine) writes: >In article <2701@wrgate.WR.TEK.COM> kend@mrloog.WR.TEK.COM (Ken Dickey) writes: >[about the Moore School machine] >>The EDVAC design was done by a group at the Moore School of Engineering >>along with von Neumann. It would not have made sense for the `first >>stored program to operate' never to have been built. >True... but EDVAC wasn't the first either!!! >The Manchester University Mark I machine was the first stored program >machine to run. It first executed a program correctly (to compute >greatest common factors) on 21st June 1948 (I believe just a matter of >a couple of days before von Neumann's team got their machine working). > Yes, but although it was perhaps less general, both these machines post-date the Colossus, which was running, decoding the German Top Secret Navy Code (amongst others?) during much of the Second World War, and therefore BEFORE 1945! Note that this is very different from the machines that decoded enigma, which couldn't really be described as stored program computers. Several aspects of Colossus were interesting, for example it had what is still possibly the fastest paper-tape reader ever built - 1000 cps I seem to remember (people will not doubt correct me if their favourate machine went faster, nevertheless, this wasn't bad for around 1942!). Unfortunately many details of Colossus are sketchy, not least because some parts are still (I believe) classified Top Secret!! (At least this certainly was the case in the mid 1980's) If anyone has some more definitive information on Colossus, I would love to hear from them! Martin usenet: ...!mcsun!ukc!minster!martin JANET: martin@uk.ac.york.minster INTERNET: martin%minster.york.ac.uk@nsfnet-relay.ac.uk surface: Martin C. Atkins Dept. of Computer Science University of York Heslington York YO1 5DD ENGLAND
albaugh@dms.UUCP (Mike Albaugh) (06/13/90)
From article <3190@husc6.harvard.edu>, by mike@mgh-znmr.harvard.EDU (Mike Vevea): > Refering to magnetic amplifiers, > In article <25079@weitek.WEITEK.COM> weaver@weitek.UUCP (Michael Weaver) writes: >>The only commercial use I have heard of was in one of the first >>core memories sold by IBM. > > I don't remember much about how they worked, but the Univac/Remington Rand > SS-80 and SS-90, built in the late 1950s used magnetic amplifiers instead > of tubes or transisters as the active element in their gates. There are > a few others around (Ed Gould?? Mike Albaugh?? Can you correct my fading ^^^^^^^^^^^^ You rang :-) > memories?). The computer club at UC Berkeley had a couple of these in > the late 1960s/early 1970s. And I personally salvaged one, but ran out of time/money in 1976 :-( Anyway, I was going to email, and still suggest email to me for followups, but there were a few points relevant to comp.arch, and I thought I'd bring them up. The mag-amps in the SS-90 (and in a similar RCA machine I know nothing else about) were not exactly as described by a previous poster (or in contemporary textbooks. Rather than having separate clock and sense windings, the "clock" winding was essentially in series with the output, hence the input to the next gate. The gates were arranged in a two-phase scheme. That is, the "A-phase" cores fed their outputs to the inputs of the "B-phase" cores and vice-versa. The basic element had only two windings. If the input winding of a core recieved a pulse, the core "flipped". On the next phase the energy of the "clock" winding was disipated in "resetting" the core, and very little appeared on the output. But if the input had not been pulsed, the clock pulse went straight through and was able to set the next stage. This is, of course, a gross over-simplification, but the basic point is that these were clocked inverters. Only the register bits were actually nearly as simple as described. A lot of "magic" went into getting usable fan-out from the more general gates. Logic was performed by diode "or" gates on the inputs of the cores. One interesting side-effect of having been exposed early to this was that I ended up taking pretty easily to NMOS, which also favors two-phase clocked dynamic storage nors :-) (while my ttl-raised companions prayed for NANDS) BTW, the card-reader contained a buffer memory implemented with "diode-capacitor-diode" storage, very much like a discrete DRAM cell! The machine was descended from the "Cambridge Air-force Computer" of 1956, but I believe there were a few in commercial use as late as 1975. Not a bad life-span... Like I said, email for details, comp.nostalgia signing off now... Mike | Mike Albaugh (albaugh@dms.UUCP || {...decwrl!pyramid!}weitek!dms!albaugh) | Atari Games Corp (Arcade Games, no relation to the makers of the ST) | 675 Sycamore Dr. Milpitas, CA 95035 voice: (408)434-1709 | The opinions expressed are my own (Boy, are they ever)
mmm@cup.portal.com (Mark Robert Thorson) (06/13/90)
I was in the computer club at UC Berkeley in 1975, and we had no hardware at that time. Later, we got two PDP-5's and after that a Honeywell (DDP-350 ?). One of the PDP-5's had a homebrew fast/slow switch on the front panel, for varying the speed of the clock. I guess you had a choice between fast answers and correct answers.
sxr@cs.purdue.EDU (Saul Rosen) (06/14/90)
> >If anyone has some more definitive information on Colossus, I would love >to hear from them! > There are several interesting articles about the Colossus in volume 5 No. 3 (July 1983) of the Annals of the History of Computing. Colossus was undoubtedly running earlier than any of the general purpose programmed electronic computers. However most people agree that Colossus was a special purpose electronic system, and not a general purpose programmed computer.
rpw3@rigden.wpd.sgi.com (Rob Warnock) (06/15/90)
In article <10814@medusa.cs.purdue.edu> sxr@babbage.cs.purdue.edu (Saul Rosen) writes: +--------------- | The Mercury Delay Line is not of real current interest in the | world of computer architecture, and there is not much point to | including a technical description here... +--------------- Well, not in and of itself. However, delay-line storage may be about to make a comeback, but instead of sound waves in mercury using light waves in air (or glass or vacuum). At least one of the research projects [e.g., at Bell Labs, per Alan Huang, several years ago at a public seminar] investigating possible all-optical computers has come up with a scheme in which the state of the "micro-engine" is stored as on and off spots in a planar wavefront of light which sweeps through the machine. Each pass through the logic block [literally a block of glass and other stuff!] corresponds to one "clock" of the micro-engine, and all of the logic gates of the micro-engine operate in parallel as the pattern of light sweeps by. Since they anticipate the switching time of the opto components to (eventually, based on device physics) be much less than the propagation time of the "wavefront" around the loop (which includes a regenerator stage), the idea has come up of stacking several micro-engine states one after another, thus making a "barrel processor" out of the beast. With a round-trip time of (say) 5ns and a switching time of 10 ps (including guard bands), you could stack 500 parallel CPUs in the same space as one. Each machine would have a 5 ns micro-cycle time. [Actually, as I recall, they were talking about switching times in the femtoseconds, not picoseconds...] By the way, current optics are good enough to handle a "wavefront" with say 100 x 100 "bits", or 1250 bytes of micro-state. Refined optics could probably handle 1k x 1k bits (?). [Can anybody who knows better comment...???] And of course, your cache memories could be additional "delay lines" which could store more bits in a longer and longer loops, until you get to speeds that could match conventional memories. And you could get a synchronous ring network connecting the various parallel "CPUs" by sending a small portion of the state "wave" through a *slightly* longer path (say, put a tad thicker glass on that portion of one of the mirrors), such that whatever bits were output by a given micro-engine during one microcycle would show up in the next engibe "back" on the next microcycle. [Hmmm... and by shaving off a few microns of glass off another area of the mirror, you could feed some bits *forward*. Counter-rotating rings!] So even though mercury delay lines are dead, delay lines may not be... -Rob ----- Rob Warnock, MS-9U/510 rpw3@sgi.com rpw3@pei.com Silicon Graphics, Inc. (415)335-1673 Protocol Engines, Inc. 2011 N. Shoreline Blvd. Mountain View, CA 94039-7311
sxr@cs.purdue.EDU (Saul Rosen) (07/07/90)
In article <645274007.10856@minster.york.ac.uk> martin@SoftEng.UUCP (martin) writes: >>The Manchester University Mark I machine was the first stored program >>machine to run. It first executed a program correctly (to compute >>greatest common factors) on 21st June 1948 >> >Yes, but although it was perhaps less general, both these machines >post-date the Colossus, which was running, decoding the German Top >Secret Navy Code (amongst others?) during much of the Second World War, >and therefore BEFORE 1945! > >Unfortunately many details of Colossus are sketchy, not least because >some parts are still (I believe) classified Top Secret!! (At least >this certainly was the case in the mid 1980's) The Volume 5 Number 3 (July 1983) issue of the Annals of the History of Computing has a special feature entitled "Colossus at Bletchley Park." It contains several excellent articles that discuss the Colossus in a great deal of detail. The cryptanalysis work for which it was built is still classified, but the machine itself is not. The Colossus was years ahead of anyone else in the successful use of electronics in a large scale special purpose computing system. The first Colossus system that was in operation in December, 1943 had 1500 vacuum tubes (valves in England.) However the Colossus was not a general purpose computer. I don't think anyone questions the fact that the ENIAC was the first electronic general purpose computer. The Colossus was certainly not a stored program computer, and it certainly did not store programs in delay line memory. Most people give Maurice Wlkes' EDSAC at Cambridge University credit for being the first stored program computer that actually ran real programs, which it did in May of 1949. The initial design of the EDVAC at the Moore School was done in 1944 and 1945, but the EDVAC did not become operational until much later. The Manchester Mark I does have a reasonable claim for being first, but it can be argued that it was a minimal computer, a prototype that could only run simple test programs. Some people at Univac have claimed that they have evidence that proves that Eckert and Mauchly's BINAC was running even earlier than the EDSAC and that it deserves credit for being the first running stored programm computer. computer that actually ran useful programs.
mmm@cup.portal.com (Mark Robert Thorson) (07/08/90)
warnock@sgi pointed out how optical computers of the future may share
a characteristic of mercury delay line computers, namely that the dynamic
state of the program is stored as one or more wavefronts in a suitable
medium. This reminds me of an idea I had for a type of fluidic computer.
A fluidic computer consists of channels cut in a solid material, through
which a pressurized liquid or gas flows. Because of certain effects which
occur near the surface of a fluid-carrying channel, it is possible to
cut shapes which act as switches. These switches have no moving parts,
except for the fluid itself.
An early backer of fluidic technology was the DoD. In the late 50's and
early 60's, they were looking for a technology to provide small flight
computers for nuclear missiles. Fluidic computers had four main advantages:
1) very rad-hard
2) cheap (they were made by etching plastic sheets, in a manner
similar to making PC boards)
3) they could be powered by pressurized gas (this is more convenient
than electricity on a rocket)
4) their control outputs are pressurized gas (again, more convenient
to control valves and flight surfaces with a cylinder than a
solenoid)
Integrated circuits knocked fluidics out of the race for the flight
computers. In so doing, they pretty much wiped out fluidic technology,
because flight computers had demands which were well-suited to fluidic's
strengths. In other applications, fluidics was a loser. The last company
to push fluidics for computation was Corning. As recently as the mid-70's,
they had a line of logic modules you could use to build small all-fluidic
digital circuits. They tried marketing them for embedded control applications
in high-noise environments, but by then transistors had left them in the dust.
My idea would be to have a fluidic computer which exists as a wavefront
in a suitable medium. The channels and switches would be cut into this
medium. The power would come from hot gases. These gases would be
products of combustion at the leading edge of the wavefront. The medium
would be like the fuel/oxidizer mix in solid rockets, and the computer would
only pass through it once (unlike the optical computer, where it recirculates).
The computer would control where combustion is most vigorous by controlling
the back-pressure at the combustion sites. By using fluidic switches to
control the release of the exhaust gases, the fluid computer can control
the rate of burning at each burning site, hence it can carve whatever
it wants into the raw unburned material ahead of the wavefront.
The computer would be implemented by casting the rocket fuel around a set
of thin wires. The computer would begin execution when a large pulse
of electricity vaporizes the wires. This would allow the initial state
of the machine to have any shape of channels and switches that might be
required.
It could proceed like the nautilus, building a new home for itself in front,
and sealing off the previous chamber behind. I would guess the initial state
would be a large version of the computer (to simplify the process of making
the exploding wires), and that each successive generation would
build a smaller version when
it replicates, until some limiting factor was reached. This strategy would
allow the designer to reach a level of resolution beyond his wire-forming
technology. To minimize the number of wires needed, the initial computer
would be sort of a bootstrap loader, which would be expecting instructions
on how to contruct the real McCoy.
Although not as fast nor as dense as electronic computers, this wavefront
fluidic computer has one property they don't have: given a suitable
medium, it exists only as its dynamic state. If it executed a halt
instruction, it would die. The only record of its existence
would be channels it had cut while it was executing.
Another unusual property is that it might exist in nature. A suitable
computing medium need only combine solid and liquid or gas states,
have a source of pressurized gas or liquid, an exhaust, and
some means for controlling the cutting of new channels. Magma recently
arrived from the earth's interior contains a lot of dissolved gas,
and at times combines the solid, liquid, and gas states. And indeed,
some examples of solidified lava contain many intricate channels, as
we might expect to find in material consumed by a wavefront fluidic
computer. This leads to the obvious question of what the minimum
requirements for a self-reproducing wavefront fluidic computer would be,
and whether such a structure could arise spontaneously, or through any
conceivable sequence of evolutionary events. Perhaps some future
civilization of computers will adopt such a theory as their creation
myth, once they've forgotten about us!aglew@oberon.crhc.uiuc.edu (Andy Glew) (07/11/90)
..> Fluidic computers [Mark Robert Thorson] >Another unusual property is that it might exist in nature. A suitable >computing medium need only combine solid and liquid or gas states, >have a source of pressurized gas or liquid, an exhaust, and >some means for controlling the cutting of new channels. Magma recently >arrived from the earth's interior contains a lot of dissolved gas, >and at times combines the solid, liquid, and gas states. And indeed, >some examples of solidified lava contain many intricate channels, as >we might expect to find in material consumed by a wavefront fluidic >computer. Rocks are just slow people! -- Andy Glew, aglew@uiuc.edu
tgg@otter.hpl.hp.com (Tom Gardner) (07/16/90)
|The last company |to push fluidics for computation was Corning. As recently as the mid-70's, |they had a line of logic modules you could use to build small all-fluidic |digital circuits. Hydraulic logic has definite advantages in hostile environments: - no electricity required, so reduced danger of explosion where flammable gases are present - robust - noise immune, so long as the pressure in the power source is in the range 2000 - 4000 psi the logic will continue to work I first came across hydraulic logic in ~1983 where it was being used to control the operation of unmanned offshore oil platforms. We concluded that (at that time) to replace the hydraulic logic with a uP would involve considerable time and expense, and that there would be no corresponding benefits...
mshute@cs.man.ac.uk (Malcolm Shute) (07/17/90)
>> Magma might form naturally occuring fluidic computers [Mark Robert Thorson] >Rocks are just slow people! [Andy Glew] But good (computer) architects. -- Malcolm SHUTE. (The AM Mollusc: v_@_ ) Disclaimer: all