Re: What ever happened to... Cray Comp/NSA co-development

[tcmay@got.net (Timothy C. May)]

Anitro wrote:

>O.K. Just a factor of 16 increase over the CM architecture right? Not knowing
>the computational capabilities of the individual processors, it might be
>difficult to say what the machine is capable of. Wouldn't that have some
>bearing?

Not in cracking "truly large" problems by brute force. Even if each of the
million processors is capable of 100 MIPS (which is unlikely, given the PIM
approach and the fine-granularity, few-bit-or-less word size, etc.), this
is only 10^8 MIPS. For problems that (for instance) 10^75 machines would
have to spend 10^10 years on, not even a drop in an ocean.

The point is that no existing machines are going to make a dent in these
calculations, though algorithmic cleverness _might_ make factoring much
easier (personally, I doubt it).


>> Second, there was work on a "processor-in-memory" architecture, in
>> conjunction with a Bowie, Maryland spook-connected company. Perhaps this is
>> what you are thinking of?
>
>I didn't mention a 'processor-in-memory' architecture and neither did the NYT
>article. Don't know about any other company involvement, just CCC and NSA.

I wrote my response before seeing the other response someone else (Thaddeus
J. Beier) gave the name of the company. Thad's details match what I recall
about this Bowie company. The archives should have articles on this, dating
back about a year or so ago. As I recall, someone on the list knew people
working at the Supercomputer Research outfit.


>> Third, all avenues of continued funding having fallen through, Cray
>> Computer (not Cray Research, of course) was shut down and assets
>> liquidated. I haven't heard what's become of Seymour, though. (He is
>> undoubtedly an asset, buy I doubt the Agency would have him liquidated.)
>
>You mean the avenues that are of PUBLIC record. The possibility could always
>exist that the development continues 'in-house'. It wouldn't be the first time
>that sort of move has been played.

You, whoever you are, asked about the operational status of Cray Computer,
and I answered. It entered Chapter Seven liquidation, parts of existing
computers were sold for scrap or museum/curio objects, the GaAs fab was
sold to another company, and the employees scattered. Claiming the NSA
moved it "in-house" is implausible, at least in terms of these events.

>> Huh? First, what evidence do you have for this claim? Second, who cares?
>> Implementing Clipper in a Cray Computer machine--why bother?
>
>[Rant mode on]
>Speculation Tim... I'm SPECULATING. Could, in theory... AND my kind of theory
>probably has holes you could drive a FLEET of Mack trucks through. I have NO
>evidence. I'm not sure WHO would care. I'M A PARANOID DILLUSIONAL PSYCHOPATH!
>O.K. well maybe not that last part... but I'm asking the questions, remember?
>I said I'm new here, so if your going to blow holes in my pet theories, then
>do me the 'kindness' of using an accurate weapon... that's why I posted...

I asked what evidence you have for this claim. Idle speculation, based on
innumerate estimates of plausibility are helpful to no one.

>> As to the claim that a million-processor machine could do this, you need to
>> work out the math. (If a backdoor exists, or the LEAF has been gotten, a
>> supercomputer is not needed....)
>
>Again... I claimed NOTHING! SPECULATED MUCH! Now it's your turn... Why would
>YOU build a machine like this? What could POSSIBLY be it's capabilities?
>Speculate with me for a moment... *_take a chance_*.

On our list, and in the circles I am familiar with, "speculations" are not
treated with more respect than are actual "claims" (whatever they are). If
you speculate that "really, really fast computers" will make today's
ciphers insecure, based on not even the simplest of calculations, expect to
be either ignored or called to task. This is the way science works, by
falsification of theories (and speculations).

>> If you mean using supercomputers to brute force the general factoring of an
>> RSA modulus, this is nonsense. While there may be math shortcuts we don't
>> yet publically know about which make factoring easier than we currently
>> think it is, a mere million or even a billion processors will not make a
>> dent in the factoring of, say, a 700-digit modulus. See the tables in
>> Schneier and elsewhere for some estimates of factoring efforts needed.
>
>Nonsense? Is that 700 decimal digits or 700 binary digits? I don't have the
>tables that you refer to. Where may I find them? (LOL)

A 2048-bit key, such as many of us use, uses a roughly 700-decimal-digit
modulus. "Schneier," mentioned in my paragraph, is "Applied Cryptography."
Page 284 of the 1st Ed. has a discussion of the work involved in factoring
large moduli. Rivest has made some more detailed estimates.

Given your rudeness and self-professed ignorance of even the basic math, I
see no point in wasting any more of my time on your posts.

--Tim May


Views here are not the views of my Internet Service Provider or Government.
---------:---------:---------:---------:---------:---------:---------:----
Timothy C. May              | Crypto Anarchy: encryption, digital money,
tcmay@got.net  408-728-0152 | anonymous networks, digital pseudonyms, zero
Corralitos, CA              | knowledge, reputations, information markets,
Higher Power: 2^756839      | black markets, collapse of governments.
"National borders are just speed bumps on the information superhighway."