Page 15 of 16 FirstFirst ... 513141516 LastLast
Results 141 to 150 of 156

Thread: pyrit CUDA nvidia Tutorial + Nvidia overclock instructions

  1. #141
    Just burned his ISO monovitae's Avatar
    Join Date
    Jul 2008
    Posts
    24

    Default

    Ok, I managed to get it working as well, albeit by a different route. This may be something some of you that are having problems may want to try.

    My original issue was that by following the original tutorial I could never get my xorg reconfigured in such a way that it would stop throwing fatal errors about the # of devices not matching the number of screens.

    So I went here >> nvidia.com/object/cuda_get.html << grabbed the 32bit version for ubuntu 8.10. Logged out of x, then ran

    root@bt:/home# sh cudadriver_2.2_linux_32_185.18.08-beta.run

    Followed the prompts in the Nvida installer, restarted X and was good to go.

    python pyrit.py benchmark returns
    439.58 PMKs/s for CPU
    1506.69 PMKs/s for Device GeForce Gt 120M

    So all seems well.

    My system if anyone cares is an ASUS N81Vg Intel Core 2 Duo P8600(2.40GHz) 4GB Memory NVIDIA GeForce GT 120M.

    If something about what I did is obviously wrong please point it out so I don't goof others up but the benchmark works, X starts, and compiz still functions so I haven't found anything wrong yet.
    ~Monovitae~

  2. #142
    Just burned his ISO
    Join Date
    Jun 2009
    Posts
    1

    Default

    Quote Originally Posted by doubleK View Post
    I cleaned it up a bit and added comments to the parts I modified (benchmark, passthrough, and split_dict).
    Here is the pastebin link:

    http :// pastebin.com/m2c420ddd
    The pastebin isn't working anymore ... does anyone have this code?

    Thanks!

  3. #143
    Junior Member
    Join Date
    Jul 2009
    Posts
    27

    Default

    Please, could somebody figure out such issue. My system is Gateway 7807 Laptop with 9800 GTS card, I have installed 185.18.08-bt5 driver, and cpyrite-cuda, then tested how it works
    root@track-laptop:~# pyrit list_cores

    The ESSID-blobspace seems to be empty; you should create an ESSID...

    The following cores seem available...
    #1: 'CPU-Core (x86)'
    #2: 'CPU-Core (x86)'

    So such driver doesn't fit to this card?

  4. #144
    smurof
    Guest

    Default

    Check code.google.com/p/pyrit/wiki/Troubleshooting

  5. #145
    Senior Member
    Join Date
    Jan 2009
    Posts
    114

    Unhappy

    last kernel 2.6.30.5 + apt-get install nvidia-drivers (190.18)
    "pyrit list_cores" crashes X, and it is not possible to use an alternate shell ctrl+alt+n
    acer 5920g , 345abg , nvidia 8600m
    bt5 kde 64bit + acpi + cuda 4.0 / nvidia 270.40 / pyrit

  6. #146
    prowl3r
    Guest

    Default

    Quote Originally Posted by Nemis View Post
    last kernel 2.6.30.5 + apt-get install nvidia-drivers (190.18)
    "pyrit list_cores" crashes X, and it is not possible to use an alternate shell ctrl+alt+n
    Just in case you didn't manage to fix it yet.

    Assuming you compiled headers, replaced symlink and created Module.symvers:

    Problem is nvidia-driver from repo is not properly recognized after installation, causing pyrit to crash. You may check this at System-Nvidia X Server Settings: "You do not appear to be using the NVIDIA X driver".

    So, a workaround for this is to get latest cuda driver from the Nvidia site.

    To complicate it just a bit more, nvidia-driver and cpyrit-cuda are mutually dependent (at the repo) so:

    - If you remove nvidia-driver, then cpyrit-cuda is removed too.
    - If you install cpyrit-cuda, then nvidia-driver is installed too.

    But we only want cpyrit-cuda installed and not that tricky driver, so:

    1. Download latest cuda driver from Nvidia site to /root
    2. Close X session and stay at console
    3. Install both pyrit and cpyrit-cuda (along with nvidia-driver) from repo
    4. Overwrite nvidia-driver with latest downloaded beta cuda driver:

    Code:
    ./cudadriver_2.3_linux_32_190.18.run
    Say yes to everything, this is:

    - Compile driver - Y
    - Try to uninstall previous driver -Y
    - Modify xorg.conf -Y

    Then apt-get update, apt-get upgrade to fix broken packages, if any.

    Startx will now show the Nvidia Beta splash and pyrit will hopefully work.

    (At least it did for me)

    Edit 1. After reading the thread again, noticed this is what monovitae already did.

    Edit 2. This procedure is not supported by the RE team. It involves installing a beta driver which may cause issues. So, basically you're flying alone and you most likely won't get support from admins / dev team should something goes wrong. The proper way is to ask admins for a fix / howto or just wait for them to update the repository.

  7. #147
    Member
    Join Date
    Jun 2008
    Posts
    50

    Default

    If I wished to run 20billion passwords on a .cap file, what would be the quickest way from start to end.

    Looking at Purehates original post it would seem the airolin db is created at 5k per sec, and the cracking then at 186k per sec. This by my calculations comes out at somewhere near 46 days from start to finish.

    Is there new hardware available or new techniques which improve on this? (please dont quote super CIA computers, I am talking hardware which is reasonable, lets say a pc for £1000 all in).

    Thanks

  8. #148
    Junior Member default's Avatar
    Join Date
    Nov 2007
    Posts
    87

    Default

    Quote Originally Posted by letmein View Post
    If I wished to run 20billion passwords on a .cap file, what would be the quickest way from start to end.
    Passthrough, creating databases is only quicker if you have caps from multiple ap's with the same SSID

    Quote Originally Posted by letmein View Post
    Looking at Purehates original post it would seem the airolin db is created at 5k per sec, and the cracking then at 186k per sec. This by my calculations comes out at somewhere near 46 days from start to finish.
    PureH@ates' airolib creation benchmark was achieved using one 8800gt GPU, I'm pretty sure you could buy more than one of them with £1000. You'd get roughly the same speed with passthrough method

    Quote Originally Posted by letmein View Post
    Is there new hardware available or new techniques which improve on this? (please dont quote super CIA computers, I am talking hardware which is reasonable, lets say a pc for £1000 all in).

    Thanks
    GPU's are getting faster and they seem to be doing about 12 PPS per GFlop. Given the GT300's are meant to hit 3TFlops, they could be doing as many as 36,000 PPS, maybe more with MIMD??

    Also, there's that scientist at NVIDIA who's predicting 20GFlops per GPU by 2015! That'll do your 20Billion passwords in about 1 day on a single GPU, knowing NVIDIA, they probably stick more than one on a card

    Then there's the untapped potential of gaming consoles, PS3's have been used in supercomputer clusters for a while now, some people are saying they could do about 2TFlop each? Pyrit can use OpenCL and aren't PS3's able to use OpenCL? anyone benched a PS3?

    edit: Apparently PS3's are only capable of about 160Gflops, there goes that idea

  9. #149
    Just burned his ISO
    Join Date
    Oct 2009
    Posts
    1

    Default

    Quote Originally Posted by doubleK View Post
    Hey, thanks for the demo! I used your trick to run pyrit on a cluster of 16 GeForce 8800 GT (source code modified to use MPI, specifically mpi4py.)

    Benchmark:
    89362.88 PMKs/s total.
    see http: // pastebin.com/f1818ef67

    Real implementation with output piped to cowpatty:
    3430600 passphrases tested in 58.11 seconds: 59034.70 passphrases/second
    see http: // pastebin.com/f41a72352

    If anyone has a cluster they can fool around with, I'd be happy to share the specific mpi code since I won't have access to my cluster anymore in a few days...

    (I can't post real URLs before I have more than 15 posts apparently... sorry about my weird URLs!)
    Hey I have access to a cluster of 40 machines with nVidia 295's at my university could I possibly have a look at you're adapted mpi code thanks.

  10. #150
    Just burned his ISO
    Join Date
    Oct 2008
    Posts
    2

    Default

    Processed all workunits for ESSID 'myessid'; 8231 PMKs per second.
    Computed 8231.39 PMKs/s total.
    #1: 'CUDA-Device #1 'GeForce GTS 250'': 6375.7 PMKs/s (Occ. 99.4%; RTT 2.9)
    #2: 'CPU-Core (SSE2)': 633.8 PMKs/s (Occ. 99.9%; RTT 3.0)
    #3: 'CPU-Core (SSE2)': 638.1 PMKs/s (Occ. 99.5%; RTT 3.0)
    #4: 'CPU-Core (SSE2)': 630.4 PMKs/s (Occ. 99.5%; RTT 2.9)

    Code:
    2830415 passphrases tested in 17.00 seconds:  166450.03 passphrases/second
    thats actually pretty nice ..

Page 15 of 16 FirstFirst ... 513141516 LastLast

Posting Permissions

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