Using the video card for breaking encryption?

[Ux]

Hi folks

There is much hype these days about hackers using video cards to break encryption.

After perusing a quick reference sheet for OpenCL, which does not have 1024-bit divisions or anything but actually offers mundane 32-bit and 64-bit operations for the most part, it is unclear to me how these hackers are using video cards for elaborate math operations that are not what the cards are intended for i.e. 3D graphics.

Are they not using OpenCL?

Can anyone offer a clue as to what the hubbub is about? Or is it just the typical computer industry hype?

Thanks.

[Bryant Keller]

Not exactly sure how this topic is related to NASM in any way, shape, or form... However, what these guys are doing is using the video card's GPU as an extra processor. I suggest looking over at wikipedia if you want more info.

http://en.wikipedia.org/wiki/GPGPU

[Ux]

Thanks for the tip. I admit it's tangential. It would be crazy to try and do GPGPU stuff from asm. But there is a connection, in that asm coders are often interested in maximizing performance.

[Rob Neff]

Maximizing performance - yes.
Breaking encryption routines - only for theorum-proving.

Please note: we're not in the habit of providing that sort of information for reverse-engineering purposes.
You'll find that stuff at other boards - not here.

[Ux]

Maximizing performance - yes.
Breaking encryption routines - only for theorum-proving.

Not everyone who asks about breaking encryption has bad intents, just as not everyone who buys a lighter is an arsonist, and not everyone who owns a knife is a murderer.

[gens]

one could argue that one who asks how to stab people ...

its not the admin's fault for presuming what they did, today rarely anyone is interested in asm for general programing

as for the question, learn opencl, it shouldnt be hard(er then ARB)

[Keith Kanios]

its not the admin's fault for presuming what they did, today rarely anyone is interested in asm for general programing

The primary purpose of avoiding the discussion of such content is to maintain a relatively high "signal to noise" ratio and relatively low burden on the moderation of this forum. It has little to do with judging someone's intent, outside of blatantly obvious/illegal stuff.

In shorter words, we do not want to attract script kiddies

A mix of GPGPU/OpenGL research and perhaps the Woodmann RCE site is the best indirect answer I can think of at the moment. Have fun.

[Bryant Keller]

Actually this took kind of an odd turn. My objection to the question was simply that it wasn't in relation to an instruction set that NASM supports. The GPGPU/OpenGL ARB instruction set is something completely separate from NASM and because of that I referred him to Wikipedia. To be honest, I don't see cryptanalysis as a bad subject to discuss at all. Of course that would be something more for a general programming or advanced mathematics forum.

[Frank Kotler]

Not too long ago, the gubmint here in the Land of the Free considered any international discussion of Public Key Cryptography "munitions dealing" - whether you encrypted anything with it or not! I think they've lightened up. Crypto that could be used to break copy-protection might run afoul of the DMCA. Better to keep the forum free of any potential legal hassles on those fronts.

As for GPU code, I'll bet macros could be written for Nasm... probably not the most appropriate tool...

Best,
Frank

[gens]

hehe i didnt know what a "french key" was until i was waiting for documents at the police station, reading their panflets
(its for breaking into cars)

as for encription, AES is becoming standard and a 8 char password(big+small letters+numbers) would keep lots of computers running hot for a looong time, even with the help of new grafics cards
(10 char would be practicaly unbreakable)

so i see no worry with people wanting to learn things like that, if they are smart enough to implement those things they should smart enough to not use them for evul (not that they could)

as for the forum noise, its the reason i made an account here, sry for helping derail the topic a bit

[Rob Neff]

AES was ratified years ago.  You can read the NIST FIPS 197 doc here: http://csrc.nist.gov/publications/fips/fips197/fips-197.pdf.
You can follow its facinating journey here: http://csrc.nist.gov/archive/aes/index.html
At this point in time I think it's safe to say it is indeed today's standard ( at least in the US anyways ).

Although I'm no crypto-analyst I do enjoy discussions regarding encryption/decryption and topics such as PKI and SSL and have implemented and used various flavors of encryption algorithms.  To expound upon my previous post in this thread my main concerns regarding public discussions of breaking encryption algorithms on this forum are mostly related to a) As Frank pointed out - potential liabilities of such discussions; and b) mis-use of that information for illegal activities.

Thus, while the topics themselves are rather fascinating and do indeed make for excellent discussions, it is simply my opinion that we let other, more specialized, forums bear the brunt of the risks associated specifically with breaking encryption algos.  Had this topic originally been about implementing encryption routines via the GPUs I think we'd have been a little less concerned.  It's a shame tho - as an open-source advocate I feel we should be able to discuss topics such as these on our forums without repercussion or threat of a potential  lawsuit.