Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
Encryption Security

New Distributed.net Clients for DESIII 47

Posted by CmdrTaco from the number-crunching dept.
David McGrath writes "The new distributed.net clients are out. If you want to participate in the DES-III contest, you should download them. I've also noticed a marked improvement for MMX processors under RC5 -- blocks take two-thirds the time to complete, at least on my computer. " It's also worth nothing that Slashdot is Neck and Neck with the Evangelista's on rc5. Keep up the good work, and remember that The odds are 1 in 4987 that we'll wrap this thing up tomorrow (grin). Anyway, upgrade and let's crush DESIII.
This discussion has been archived. No new comments can be posted.

New Distributed.net Clients for DESIII More

New Distributed.net Clients for DESIII

Comments Filter:
  • I don't see how they *can* check for this. They give out keys, and then the client reports that it's either found a match or it hasn't. How can distributed.net know if it's actually checked the keyspace or just reported something falsely? Maybe that's why the protocols and source are all closed - in this case, security through obscurity is the only possible security. If all the protocols and the source were open, then somebody could also modify the client so if it found the key, it'd tell the person running the client and not report it to the distributed.net servers, so the guy could claim all the RSA prize money for himself, instead of just getting the 10% that distributed.net gives to the person who found the key.

    Security through obscurity is usually a bad idea, but in this case, I can't think of any way to really make this secure other than making it obscure.
  • Heh, oops, hadn't thought of that. I suppose that would solve both the "not checking blocks" problem and the "keeps solution to himself" problem. They'd just have to make sure they do enough random checks to keep all the cheaters at bay.
  • Hi

    Linux accounts for more than twice as much cracked keys per day as Mac!
    And it's share is growing, too. Yesterdays shares per OS:

    Windows: 66.1%

    Linux: 14.0 %

    MacOS: 6.2%
    Have a look at this [univie.ac.at] graphs for details.

  • I was thinking more along those lines too. Celeron prices are going down so you can get better deals now.

    If you like the PowerPC, you can cluster about 20 6100/66's on top of each other for about $250 @. For about $5,000.

  • Distributed.net's clients aren't open-source, so patches are out. A frontend, though, that spawned rc5des and grabbed its stdin and stdout could be open-source (and could compile and run on any *nix system that Windowmaker does). The issue, of course, would be whether a frontend for a non-free program could be considered free.
  • I know that Rob is a minor deity in many people's eyes... but...

  • and it does seem like something the l00ft would do. Does distributed.net check for this kind of garbage?

  • It is pretty easy for even a d.net critic like me to see that v3 is what they were talking about. The latest clients are _still_ just extensions of the security-through-obscurity v2 stuff. Apparently my last jab did get Beberg to update his finger page, so at least someone is listening(Nuggett I assume). Whether anything is actually getting done or not is another question.

    I know that I'm not going to bother with this upgrade. It's just too much trouble for the return. Perhaps I'll put it on the two machines that I use every day(after waiting a couple of weeks for the bugs to sift out), but it's just too much trouble to deal with fifty-odd boxes and their users for basically no return.

    If they don't run DES-3 or the test run, so what? They'll still crank on RC5-64, no gain, no loss, no motivation, no effort.

    Seeya when v3 arrives, unless SETI@Home arrives first.
  • ...they don't update the web page very often, but there is a live mailing list and they have gotten the funding that they needed.

    Hell it better still be alive, I contributed some source(bug fixes and screen output formatting) a little while ago!

    They are supposed to start up in April. Last I heard that schedule is still valid.

    Hmm, I'll send a note and see for sure...
  • Are _so_ much better at DES because they are 64 bits wide and can take advantage of the more efficient bitslicing algorithm.
  • We need more people to push the total keyspace that the Evangelistas have already done and to keep L0pht Heavy Industries off our backs. L0pht is accelerating at an alarming rate. Just a short while ago they weren't even in the top ten.

    Look at this delta [distributed.net]

  • You used to be annoying, now you're just amusing.

  • It looks like if you follow the links, you end up getting a 2.7100 or 2.7102 client. But if I'm reading correctly, you need the 2.7103 client to do DES III. Try getting the latest client from here [distributed.net]. It says libc5, but it really isn't linked to any libraries at all, so it's okay to use no matter what libc you have.
  • A good description of what you need to do [slashdot.org] was posted on Slashdot a few weeks ago as a story.

  • EFF?? (Score:1)

    by asn ( 4418 )
    Does anyone know if EFF is going to build a bigger/better DES Cracker to go after the DES-III contest?
  • Slashdot is nowhere _near_ the Mac team -- I'm not dissing any team but Team Slashdot has submitted about HALF as many blocks as the Evangalista's. These "leads" being posted are of a few percent, and mostly temporary.

    (Yes I'm not on a team but that's nothing to do with it).

    Regarding the lack of an open sourced client.. can't someone just reverse engineer the app? Should be simple to do, no?

    Scott
    Linux user, Mac user, Windows victim.
  • Under windows NT you can configure the client to install as an NT Service. It will continue to process after you have logged out. See the readme for details..

  • Question (Score:1)

    by Chas ( 5144 )

    A buddy of mine has tried the client on both a K6-2 300 and 333. It'll run (well, walk) the client with nothing else running. Unfortunately, if you're doing anything else, you take a big performance hit and your keyrate dogs out.

    He actually locked the 300 machine twice doing nothing more than RC5 and IRC.

    Use whatever you have, just be aware.

    What am I running? Currently I'm running a P2 333 with 192 MB RAM and about 9 gigs of disk space. I'll be upgrading in a month or so to a CeleronA OC'ed to 450 or 504. Total cost? $500.


    Chas - The one, the only.
    THANK GOD!!!

  • I'm running a mac, and linux, on a PPC box, and I'm on the /. team.

    Mac's are by far not as evil as your NT machines.
  • My G3 kicked ass in the DES-II, under macos and linux, doing about 1.3 megakeys a second.
  • First, if it takes advantage of MMX, then it will also take advantage of the K6 MMX, which is cloned from Intel.

    Second, the FPU is irrelevant for encryption. Encryption has to be exact, and floating point numbers are exact only up to a certain precision. Encryption software frequently uses bignums, which are integer but can be very long.

    Probably the program makes use of the FPU for some optimization, but that's a)unlikely (at least I can't think of useful application of the FPU in encryption) and b)wouldn't be a hard impact on the performance as encryption is integer based.

    Anyone hit me if I'm wrong. But, please, not too hard.
  • Soon, there will be another way to spend those extra CPU cycles: the Search for Extraterrestrial Intelligence (SETI) is designing a distributed client to search through astronomical data for signs of something that seems somewhat lacking here on Earth (at least in the USA).
  • Ummm Version 3 is going to be open source.... The website has stated that obviously. The clients are closed source so that blocks are not faked.


    Ex Machina "From the Machine"
    ExMachina@GeekMafia.dynip.com [http://GeekMafia.dynip.com/]
  • If I run the client on a Windows machine, and then log out, does it stop running, or does it continue in the background like unix??
  • I run the program on half a dozen K5 machines, 2 Cyrix machines, a couple K6-2's, and a couple Pentium machines and have never seen any performance hit from any of them running at the least nice level. All these machines are used for intensive real world use too. If you're getting locked you must be doing something wrong.
  • I wish the Distributed.Net software was more powerful. If it could be set to auto-upgrade itself (possibly even be remote admin'able) and run any math-related applet (not Java, ewww) sent at it that'd be cool. I'd love to see an open source system similar to this be added into distributions of Netscape and ICQ. All they'd need to include would be a tiny program to check for the software the first time XYZ software ran the installation ware could ask the user if they wanted to join the collective and if so d/l and install the dist ware too. No big deal. Someone beg this into Mozilla's installation! ;)
  • Is there any such listing? I have 4 processors running for the Slashdot team...is that tracked anywhere? I'd be interested to see how many processors/people are on each team at any given time as that might put into perspective

  • True, but don't stats start over after DES-III, like they did with DES-II? Meaning that any close lead you have to them will be lost after the DES-III contest. Macintosh Computers have PowerPC processors, not suited for DES-III, but for RC5. I know a lot of Mac Users just skip the DES and save up the RC-5 blocks, waiting for the contest to end (in this case, >=2 days) and just let loose. That'll basically give them an advantage.

    Oh, and if you think MacOS isn't for those of us who like to hack (in the old school reference) then deal with this fact: We're #1, and 36% of the Blocks submitted from D.net are macs....

    Now if Apple would just bundle the D.net Client with the iMac.........
  • I was just wondering if anyone could help me out on how I would join the slashdot team? Thanks alot.
  • I will soon have 20 processors on the project (okay, so most of them are only 486's)... There's lot's of talk about joining the /. team, and I'd do it -- but I've searched high and low and can't figure out how. Anybody?
  • I see that Win, Mac, Linux, and FreeBSD all have new clients, but does anyone know when the updates for other platforms will be available. My x86 Solaris box at work is ready for a new client.

Slashdot Top Deals

Brain off-line, please wait.

Close