Results 1 to 12 of 12
  1. #1
    Silver Lounger lumpy95's Avatar
    Join Date
    Feb 2013
    Location
    Mojave Desert CA
    Posts
    2,004
    Thanks
    293
    Thanked 197 Times in 162 Posts

    Nvidia’s GTX 1080 can crack passwords as easily as it can game

    Aren't you glad they figured this out and announced it.
    https://www.yahoo.com/tech/nvidia-gt...234732303.html

    Making a crack at a password-protected Microsoft Word document, Nvidia’s GTX 1080, designed using the company’s new Pascal architecture, can reportedly make 7,100 different password attempts per second, while even the last generation GTX 980 can only make around 3,800 attempts per second, and a CPU alone can only make around 30 attempts per second. This is according to testing done by Oleg Afonin from security company Elcomsoft.

  2. #2
    WS Lounge VIP
    Join Date
    Dec 2009
    Location
    Earth
    Posts
    8,517
    Thanks
    54
    Thanked 1,038 Times in 966 Posts
    That is why a longer password is better than a shorter password.
    https://www.grc.com/haystack.htm

    cheers, Paul

  3. #3
    Silver Lounger lumpy95's Avatar
    Join Date
    Feb 2013
    Location
    Mojave Desert CA
    Posts
    2,004
    Thanks
    293
    Thanked 197 Times in 162 Posts
    Yup, it all depends on how bad they want your password. The video card just cuts down their time in finding it.

  4. #4
    Star Lounger
    Join Date
    Jul 2013
    Location
    Murphy, NC
    Posts
    74
    Thanks
    0
    Thanked 8 Times in 8 Posts
    The headline strikes me as hype. By my rough calculation, to test all passwords made up precisely of eight lower case letters would still take more than 100 days of computing.

  5. #5
    Super Moderator
    Join Date
    May 2002
    Location
    Canberra, Australian Capital Territory, Australia
    Posts
    5,121
    Thanks
    2
    Thanked 439 Times in 362 Posts
    That's just slightly longer than 3 months. In all likelihood, the average would be half that. Well worth the effort in many cases. Besides which, one could employ a divide & conquer approach with multiple PCs working away on different combos.
    Cheers,

    Paul Edstein
    [MS MVP - Word]

  6. #6
    Star Lounger
    Join Date
    Jul 2013
    Location
    Murphy, NC
    Posts
    74
    Thanks
    0
    Thanked 8 Times in 8 Posts
    The password grammar in my example wasn't especially robust. Including upper case letters, digits, and punctuation would increase the number of possible passwords by several orders of magnitude. Sure, a battalion of PCs could be employed in a cracking attempt. But the need for this is hardly suggested by the words, "can crack passwords as easily as it can game."

  7. #7
    Super Moderator
    Join Date
    May 2002
    Location
    Canberra, Australian Capital Territory, Australia
    Posts
    5,121
    Thanks
    2
    Thanked 439 Times in 362 Posts
    Quote Originally Posted by DavidHLevin View Post
    the need for this is hardly suggested by the words, "can crack passwords as easily as it can game."
    Last time I played a game on my PC, it was probably solitaire... I doubt a GTX 1080 would have been taxed - plenty of time for password cracking between moves...
    Cheers,

    Paul Edstein
    [MS MVP - Word]

  8. #8
    Star Lounger
    Join Date
    Jul 2013
    Location
    Murphy, NC
    Posts
    74
    Thanks
    0
    Thanked 8 Times in 8 Posts
    How about we agree to disagree?

  9. #9
    Super Moderator satrow's Avatar
    Join Date
    Dec 2009
    Location
    Cardiff, UK
    Posts
    4,670
    Thanks
    305
    Thanked 604 Times in 503 Posts
    The GeForce GTX 1080 only officially supports two-way SLI, which means that you could use 2 of them for cracking, with much reduced times. The older GTX 980Ti (and most other GeForce GPUs) can utilise up to quad SLI (AMD owners have similar CrossFire capabilities).

    You could use multiples of these machines in a cluster, reducing the time even further.

    You could also rent some time on one or more (of Amazon's?) servers/clusters or rendering farms, if you wanted to use your own rig(s) for gaming.

  10. #10
    Star Lounger
    Join Date
    Jul 2013
    Location
    Murphy, NC
    Posts
    74
    Thanks
    0
    Thanked 8 Times in 8 Posts
    After further pondering the headline, I think that my fundamental concern was that the phrase "as easily as" seemed misleading or nebulous. Perhaps the phrase "as well as" would have been a better choice.

  11. #11
    2 Star Lounger
    Join Date
    Aug 2014
    Posts
    118
    Thanks
    0
    Thanked 3 Times in 3 Posts
    Quote Originally Posted by DavidHLevin View Post
    After further pondering the headline, I think that my fundamental concern was that the phrase "as easily as" seemed misleading or nebulous. Perhaps the phrase "as well as" would have been a better choice.
    That would be no good. It wouldn't cause a bunch of people to go and tell everyone that will listen that their password are history and the world will end if they don't use 200 character passwords with 100 non-printing characters. Which as far as I can figure out is the purpose of headlines.

  12. #12
    5 Star Lounger Lugh's Avatar
    Join Date
    Jun 2010
    Location
    Indy
    Posts
    657
    Thanks
    181
    Thanked 80 Times in 70 Posts
    Quote Originally Posted by DavidHLevin View Post
    The headline strikes me as hype.
    Definitely click bait.

    I recall reading a few years ago that some scientists were using clusters of GPUs for certain number crunching tasks. Their narrower focus makes them pretty powerful for suited tasks apparently.

    ETA nVidia article, 2013 and PC Mag article about ATI, 2006.

    Quote Originally Posted by NVidia
    These are the same CUDA features available on the latest supercomputers like the Titan system at Oak Ridge National Labs and Blue Waters at NCSA, which make them ideal for parallel programming
    Quote Originally Posted by PC Mag
    using the massively parallel-processing power of modern graphics cards to perform certain non-graphics tasks. Some types computing problems require crunching numbers (usually floating-point math) on huge sets of data. Some of these require lots of branching logic, others can be handled in more of a streaming fashion. The latter type of task is the perfect candidate for GP-GPU (general-purpose computing on GPUs) acceleration.
    Last edited by Lugh; 2016-08-26 at 03:04.
    Lugh.
    ~
    Windows 10 Pro x64 1607; Office 2016 (365 Home) x32; Win Defender, MBAM Pro

    ASRock H97 Anniversary; Xeon E3-1231V3 (like i7)
    Gigabyte GeForce GTX 970; 12GB Crucial DDR3 1600
    Logitech MX Master mouse; Roccat Isku kb

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •