Colossus has been Rebuilt

Max Driver writes "In celebration of D-Day, "Colossus", one of the earliest electronic code-breaking machines, has been rebuilt after ten years of effort by computer conservationists. Colossus was used to break the Lorenz cipher. This story is being reported by the BBC. Remarkably, the use of parallel processing (five tape channels) and short gate delay time (1.2 microseconds) allows the Colossus to match the speed of a modern PC."

good design

[millahtime]

Remarkably, the use of parallel processing (five tape channels) and short gate delay time (1.2 microseconds) allows the Colossus to match the speed of a modern PC."

This definitely shows you what a good design can do. WIth all the advancement I expected that thing to be slower than my TI-89 calculator.

Re:Free information.

[meringuoid]

If information about this machine had been made public in the years after the war, we may now have been a good few megahertz ahead of our selves in computer technology.

I seem to remember hearing that a lot of Third World countries carried on using the German cryptosystems for a long time after the war, and that was why all the Bletchley technology was kept black - we rather liked being able to read everyone's mail. Don't know how true that is, though...

IIRC, GCHQ also invented the RSA cipher years before it was discovered in the civilian world. Damn shame we didn't get to cash in on that one :-)

Go and visit Bletchley Park!

[salmacis2]

Colossus, and indeed the rest of the Bletchley Park operation was a tremendous example of war-time ingenuity.

I would urge all UK-based \.ers to go and visit Bletchley Park as soon as possible. It's an amazing day out. It's just sad that the UK government doesn't appear to recognise the historical significance of BP and spend whatever is required to restore the site. Hut 6 and Hut 1, where most of the decoding was done are practically falling down these days.

Re:A modern PC could emulate it in physics!

[FraggedSquid]

Quite a few years ago there was an interview with one of the guys who worked on the Colossus. He stated that he had produced a machine code implementation of the task and ran it on the best PC he could find (may have been a PI or PII), expecting the PC code to run faster. He was surprised to find that Colossus was still much faster.

Re:Not really

[rosbif]

Ah, the Dick Van Dyke theory of England (for those not aware, DVD played a cockney chimney sweep in the film "Mary Poppins", thereby setting back UK-USA relations a few decades)
1. Most people in England do not have computers based on 486s - I'd be surprised if it was more than 10%. I would suggest that low end P4s are in the majority
2. Any PC with a serial port can read a paper tape with a suitable paper tape reader attached (I've done this in the recent past)
3. Better than a kid whinging on about things that he doesn't understand.

BTW, I think you'll find "mates" to be an Aussie soubriquet, rather than an English one.

Re:A tragedy

[meringuoid]

Apparently, all the parts that went into making the beasties was "borrowed" from British Telecom. After the war, they just gave the parts back.

Reminds me of something I heard about the Manhattan Project, which was a similar exercise in rounding up every geek in the country and making them do cool secret stuff... Apparently they couldn't get the copper wire they needed for the electromagnets used in refining their uranium, so they just took all the silver out of Fort Knox and made it into wire. Melted the lot down after the war and put it right back, no harm done...

Of course that makes me wonder what Auric Goldfinger was thinking of. America's loot stash is already radioactive! :-)

Re:Free information.

[FraggedSquid]

At the end of the war, the British Gov gave Enigma machines out to various bits of the Commonwealth to use. I think that this is mentioned in Simon Singhs The Code Book, but I could be wrong.
Nobody in the outside world, not even the Germans(though some has suspected), knew that Enigma had been broken till IIRC the 1960's.

Re:Free information.

[ezzzD55J]

I seem to remember hearing that a lot of Third World countries carried on using the German cryptosystems for a long time after the war, and that was why all the Bletchley technology was kept black - we rather liked being able to read everyone's mail. Don't know how true that is, though...

Well, there is something related here; Dennis Ritchie dabbles in cryptography [bell-labs.com]. He talks about cryptanalysis of the hagelin m-209b [iacr.org] crypto device (I bought one on ebay :)). They submitted their findings for voluntary review by the NSA before publishing, and Ritchie was visited by a "Retired Man" from the NSA. The relevant bit:

He got a bit more specific about two things: the agency didn't particularly care about the M-209. What they did care about was that the method that Reeds had discovered was applicable to systems that were in current use by particular governments, and that even though it was hard to imagine that these people would find the paper and relate it to their own operations (which used commercially-available crypto machines), still... perhaps we should exercise discretion? It was certainly legal to publish, but publication might cause difficulties for some people in the agency.

Full story in the first link.

So, even though this has nothing to do with the UK and colossus/enigma/lorenz directly, it still is a similar story.

Re:A tragedy

[jdtanner]

Not the only thing us Brits have missed out on I'm afraid... The integrated circuit [bbc.co.uk] RSA encryption [theregister.co.uk] Doh!

Re:Wait! Wait! there's a pattern here

[Moraelin]

Really? I thought it was the millions of Russians who died. The Americans got anywhere _near_ the war after the Russians were already stopping the Germans.

And those strategic bombings never did much damage either. In fact, it cost the US far more to bomb Germany, than it cost Germany to rebuild the odd factory that got hit by a bomb and replace/repair the fighters.

Now I'm not saying that US didn't help, and we're all grateful for that. (If nothing else, otherwise the whole Europe would have ended up communist.)

But, no offense, claiming to basically have singlehandedly won the war is a tad shameless. Without the USSR to hammer the Germans from the other side, and without the UK as a base, the US wouldn't even have made it onto the European mainland. Much less beatten Germany.

Brit RSA encrytion

[BlightThePower]

For more information see "The Code Book" by Simon Singh.

It was developed by the superbly named Clifford Cocks, a at GCHQ in 1973 (IIRC thats three years before Rivest et al.) Apparently he thought it no big deal (completing an implementation of Ellis' original proof-of-concept practically overnight) and filed it away for further reference. End of story. Cocks is now chief mathematician at GCHQ; and given that he's probably intercepting this communication as I write, I dare say he will pop-up if what I've said is inaccurate!

The true tragedy is obviously that RSA is called RSA, rather than the far more amusing "Cocks Encryption" or similar. The sheer weight of punnage (e.g., "Hard Cocks Encryption" anyone?) is a tragic is a loss to humanity IMHO.

Re:Let the british have their moment in the sun

[Shmooze]

Ah well, since no-one can accept british superiority, here is an abridged list of british inventions ripped of by the good ol' US of A.

1) Computers
2) 'RSA' encryption
3) Jet engines
4) All-Moving tailplanes (to allow supersonic jets)
5) Jump-jets (namely, Harrier)
6) Radar and Microwave ovens
+ Many more but i'm feeling far too lazy.

Re:Free information.

[raygundan]

Meh. That this device is "as fast as a modern PC" for the single task it was designed for is nothing particularly interesting. It is, as another reader pointed out, essentially just a large DSP. Just because a 400MHz GPU is many, many times faster than your 3GHz CPU at drawing pretty pictures, doesn't mean it's a better general purpose CPU. If you took all the millions of transistors in a P4 and made them all do NAND in parallel, you would have the world's fastest NAND gate, capable of doing a million near-instantaneous NANDs simultaneously.

This is not to belittle the achievement of the folks who built Collosus. It is, however, more correctly compared to things like the EFF DES key cracker, which like the Colossus was massively parallel. It was also a gajillion times faster than a PC *at the one thing it did*.

Re:A tragedy

[Isofarro]

If the British Government hadn't been so short-sighted, the UK now would be the centre of the global computer industry.

From the article, to get around the reliability of valves the solution with Colossus was to leave it on until the end of the war, so it would have been on from 1 February 1944 through to at least 15th August (surrender of the Japanese). That's a 18 month uptime.

More uptime than the average Windows laden PC.

Re:A modern PC could emulate it in physics!

[Anonymous Coward]

from a Britsh Telecom website (www.connectedearth.com):

"Colossus was then built to find the Lorenz wheel settings used for each message, using a large electronic programmable logic calculator, driven by up to 2,500 thermionic valves. The computer was fast, even by today's standards. It could break the combination in about two hours - the same as a modern Pentium PC. " ...Meaning it was as fast as a reasonable pentium at running the algorithm to crack lorenz code wheel settings, which is a perfectly reasonable statement.
Do not take this to mean it's Quake III frame rates were any good.

Re:Let the british have their moment in the sun

[garyok]

I would have added

7. Inalienable human rights (Magna Carta)
8. Liberal democracy (John Stuart Mill, John Locke, etc., etc...)

but the Americans don't seem to be using them any more. Can you send them back to Britain please if you're finished with them please?

Re:Reminder:

[Jeremy Erwin]

The quartz compositor treats each window's content, and each window as a openGL primitive. All Quartz Extreme requires is a certain amount of VRAM--32 is preferred, and the ability to support textures of arbitrary (not powers of two) height and width. As the Mac only supports a small number of video cards, this practically guarantees that a GPU will be available.

But the GPUs in early nVidia and ATI cards are fixed function anyway-- useless for all except computing Transform and Lighting. Later models (GeForce3, Radeon 8500) were programmable, but did not fully support floating point math. The latest two generations can theoretically be used for general purpose computing, but this is experimental, and only applicable to certain classes of computation.

Re:The Atanasoff-Berry Computer is slighted again

[pmc]

Here we go again. There are four contenders in this race "First {suitably qualified} Computer":

The Zuse Zn (Z1 - 1938, Z2 - 1939, Z3 - 1941)
Colossus (1944)
ABC (between 1938 and 1942)
Eniac (1946)

The ABC was not Turing complete (and, indeed, not programmable), and was probably beaten into production by the Z3 anyway . The other three are Turing complete. The Z3 was the first to be Turing complete, but it was only realised in 1998 that this was true. Colossus was Turing complete (and this was known at the time - Turing worked next door after all), but was classified top secret. Eniac was Turing complete too, but was definitely last.

So, depending on your definition of computer and how "electronic" you insist on it being, you can pick any of them. But in my opinion the ABC has probably the narrowest of claims, with Colossus the best claim. Eniac definitely had the greatest influence.

The "electronic" part of this argument is in my opinion a complete red herring. Imagine in a few (?) years time when (if) nano-technology comes of age and instead of electronic switches we go back to nano-mechanical switches. Are these computers somehow inferior to the ABC just because they are once more electromechanical like the Zuse Z3 (albeit with switches a billion times smaller)? No, I don't think so.

This is not to criticise the ABC though - it was an impressive acheivement in its own right. But too much is claimed for it: for example it was not the first to use binary as is claimed - the Z1 used binary.

Related stuff

[jarek]

Some of the stuff in the links below will be found in the Code Book. It's interesting stuff anyway.

link1 [zen.co.uk]

link2 [zen.co.uk]

Happy reading.

/jarek

Re:Wait! Wait! there's a pattern here

[naily]

Hear hear!

I'm waiting for 'Colossus' the movie, starring a daring team of americans at Bletchley who single-handedly invent Colossus, run Ultra and crack the codes just in time, all the while undermined by those stuck-up brits who always try to spoil everything by saying "You bloody yanks can't just storm in here and expect to win the war in a week!".

An old ex-empire Britain may be, but they were the first empire to dismantle itself (for the most part), and every territory they lost was made, by and large, in their image. The english civil war was over 50 years before the french revolution - they created the first democracy in europe (by killing their king - can you imagine usurping your own president?), and the american ideals of democracy were descended from this. Oh, and industry? The brits invented that too - Adam Smith was no yank!

My point is that every empire has its day, even with the best will in the world. And when your time comes you can either acceed to the rising benevolent power (US, not Nazi), or have power wrestled from you.

Re:A tragedy

[PapayaSF]

You're almost right. They did get tons of silver (not all of it) to make electromagnets (not just wire), which were so huge and powerful that when turned on, people standing many yards away could feel the pull on the nails in their shoes and on their belt buckles!

Re:Wait! Wait! there's a pattern here

[DunbarTheInept]

It's not entirely true that the bombing didn't have much damage. The reason the German industry grew immune to it was that the Germans would rebuild factories further to the east, to escape the range of the bombers. But that had the direct effect of putting the factories closer to the Russian Front - which wasn't exactly a safe haven either. Because of Russian advances on the ground, the germans did lose industry very quickly in the end and thus lose their ability to supply their armies and thus they collapsed rather fast once they were back within their own borders. But it was allied bombing that forced them to put the factories where the Russians could overtake them, instead of keeping them safely in the core of Germany. So, yes, the allied bombing had a very big effect on the war - it forced the Germans to disperse their industry to the periphery, and thus it sped up the ending phase where the Germans were in retreat.

It's still a lie to say the US is solely responsible, of course. I agree with you on that sentiment. (Although it's important not to downplay the Pacific theatre, in which the US was in a position to do most of the effort, and did so despite putting less resources into it than into the European theatre - good cryptography played a major role there - the US knew the Japaneese codes and therefore could predict exactly when and where to concentrate forces to handle Japanese attacks, and thus could beat the japaneese even with the drastically reduced forces left after Pearl Harbor.)

No 5 is (almost) alive ~ CSIRAC

[goon]

CSIRAC [wikipedia.org] - (1949 - 1961) - digital computer, entire machine housed at melbourne museum [vic.gov.au] (victoria, australia) after service with CSIRO ( formerly called CSIR), Radio physics lab Sydney University finally residing at Melbourne University [mu.oz.au].

Interesting facts [mu.oz.au] ...

• approx 5th digital computer created
• 
  
  one of last original computers intact
  CSIR Mk1 or CSIRAC designed by team lead by Maston Beard and Trevor Pearcey for CSIR (CSIRO [csiro.au])
  primary store of 768 20-bit words
  magnetic drum 4,096 word capacity
  10ms access time
  clock speed 1000Hz
  serial bus
  paper tape input
  30 KW power requirement
  crt output of registers
  high level programming via language INTERPROGRAM
  audio output for errors
  first computer programmed for music [abc.net.au]
  emululator available [mu.oz.au]
  

references:
http://en.wikipedia.org/wiki/CSIRAC [wikipedia.org]
http://www.cs.mu.oz.au/csirac/csirac.html [mu.oz.au]
story on recreations of some of the original music tracks CSIRAC [abc.net.au]
50th Anniversary of the CSIRAC [abc.net.au]