Results 1 to 6 of 6

Thread: Pyrit Not Seeing Gpu Cores

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

    Default Pyrit Not Seeing Gpu Cores

    Ok this may be a bug or more likely that the process is just different than it was in BT4 Beta.

    In beta if I were to grab the CUDA driver from
    hxxp://xxx.nvidia.com/object/thankyou.html?url=/compute/cuda/2_2/drivers/cudadriver_2.2_linux_32_185.18.08-beta.run

    End my KDE session and do

    sh cudadriver_2.2_linux_32_185.18.08-beta.run

    then follow all the prompts followed by a reboot

    Pyrit would recognize my gpu core and get roughly 1500pmk on a gt120m.

    In pre final following the same steps results in pyrit benchmark returning about 300 pmk. And pyrit list_cores not showing the gpu.

    Is there something I am just forgetting or is this an issue. WIll explore further tonight and if solved will post solution.
    ~Monovitae~

  2. #2
    Junior Member Zermelo's Avatar
    Join Date
    Feb 2010
    Posts
    54

    Default

    I would suggest downloading pyrit from svn, then compiling "cpyrit" and "pyrit" and building both.

  3. #3
    Just burned his ISO monovitae's Avatar
    Join Date
    Jul 2008
    Posts
    24

    Default

    Ok got it working. Thanks to Zermelo for the nudge in the right direction. As follows is a mini how-to from start to finish of how I got cuda running feel free to move it to How-Tos if its better there.

    So given a fresh BT4 Pre-Final first lets gather some things we need.
    First is the nvidia cuda driver and toolkit.
    Both can be found here hxxp://xxx.nvidia.com/object/cuda_get.html just choose linux 32 bit and ubuntu 8.10 then download the driver and the toolkit.

    In my case I downloaded to root so log out of X and type
    Code:
     sh cudadriver_2.2_linux_32_185.18.08-beta.run
    Follow the prompts the important ones are to overwrite existing drivers, not to download from ftp, and to compile, and finally to let it auto write your xorg.

    Once thats done lets install the cudatoolkit you normally don't need this but to get pyrit working I had to compile from source so this IS needed.

    Code:
    sh cudatoolkit_2.2_linux_32_ubuntu8.10.run
    Just follow prompt and accept default paths pretty straight forward here.

    Then you need to add the new cuda files into your path environmental or you can check to see if they are there magically by
    Code:
    cat $PATH
    If not lets do
    Code:
    PATH=$PATH:/usr/local/cuda/bin
    Then the toolkit install recommends that you update your LD_Library_Path by
    Code:
    nano /etc/ld.so.conf
    paste /usr/local/cuda/lib into the file and overwrite then do
    ldconfig
    Ok now that we have the prerequisites out of the way lets get on to the pyrit side of things.

    Grab pyrit and cpyrit from the website or feel free to use SVN etc to get the latest builds.
    Code:
    wget hxxp://pyrit.googlecode.com/svn/tags/0.2.2/Pyrit-0.2.2.tar.gz
    wget hxxp://pyrit.googlecode.com/svn/tags/0.2.2/CPyrit-CUDA-0.2.2.tar.gz
    Navigate to your download directory and extract your downloads.
    Code:
    tar -zxvf Pyrit* 
    tar -zxvf CPyrit*
    cd ./Pyrit*
    python setup.py make
    python setup.py install
    cd ..
    cd ./CPyrit*
    python setup.py make
    python setup.py install
    Assuming you have no errors that should do it. Lets check it out.
    Code:
    root@anubis:~/CPyrit-CUDA-0.2.2# pyrit list_cores
    Pyrit 0.2.2 (C) 2008, 2009 Lukas Lueg hxxp://pyrit.googlecode.com
    This code is distributed under the GNU General Public License v3
    
    The ESSID-blobspace seems to be empty; you should create an ESSID...
    
    The following cores seem available...
    #0:  'CPU-Core (x86)'
    #1:  'CUDA-Device #1 'GeForce GT 120M''
    
    
    root@anubis:~/CPyrit-CUDA-0.2.2# pyrit benchmark
    Pyrit 0.2.2 (C) 2008, 2009 Lukas Lueg hxxp://pyrit.googlecode.com
    This code is distributed under the GNU General Public License v3
    
    The ESSID-blobspace seems to be empty; you should create an ESSID...
    
    Running benchmark for at least 60 seconds...
    
    CUDA-Device #1 'GeForce GT 120M': 1490.69 PMKs/s, 93.03% occupancy
    CPU-Core (x86): 199.47 PMKs/s, 99.32% occupancy
    
    Benchmark done. 1690.16 PMKs/s total.
    Quite a bit better than my original 299 PMKs/s with just the p8600 dual core. Hope this helps anyone with similar issues. If anyone sees anything I goofed up let me know and I will edit this.
    ~Monovitae~

  4. #4
    Developer
    Join Date
    Mar 2007
    Posts
    6,126

    Default

    I'm not sure why I even bother building this crap. Every single one of those things are built as .debs and present in the back track repo.
    Code:
    root@backtrack:~# apt-cache search cuda
    cuda-multiforcer - cuda development tools
    cuda-gdb - cuda debugger
    cpyrit-cuda - Nvidia-CUDA core for Pyrit
    cuda-sdk - cuda code samples
    cuda-toolkit - cuda development tools
    root@backtrack:~# apt-cache search pyrit
    pyrit - GPU-accelerated attack against WPA-PSK authentication
    cpyrit-stream - Pyrit module for ATI Stream
    cpyrit-cuda - Nvidia-CUDA core for Pyrit
    pyrite-publisher - Convert html and text documents to palm DOC format
    root@backtrack:~#

  5. #5
    Senior Member kidFromBigD's Avatar
    Join Date
    Jan 2010
    Location
    Texas
    Posts
    159

    Default

    Quote Originally Posted by monovitae View Post
    Ok got it working.
    Same here.
    But...
    Code:
    tar -zxvf Pyrit* 
    tar -zxvf CPyrit*
    cd ./Pyrit*
    python setup.py build
    python setup.py install
    cd ..
    cd ./CPyrit*
    python setup.py build
    python setup.py install
    Those changes make the process a bit smoother.
    Quite a bit better than my original 299 PMKs/s with just the p8600 dual core. Hope this helps anyone with similar issues. If anyone sees anything I goofed up let me know and I will edit this.
    My output:
    Code:
    The ESSID-blobspace seems to be empty; you should create an ESSID...
    
    The following cores seem available...
    #0:  'CPU-Core (x86)'
    #1:  'CUDA-Device #1 'GeForce GTX 260''
    And...
    Code:
    Running benchmark for at least 60 seconds...
    
    CUDA-Device #1 'GeForce GTX 260': 8118.49 PMKs/s, 99.13% occupancy
    
    Benchmark done. 8118.49 PMKs/s total.
    Thanks!
    You. Are. Doing. It. Wrong.
    -Gitsnik

  6. #6
    Developer
    Join Date
    Mar 2007
    Posts
    6,126

    Default

    Let me rephrase my self. If your bug or fix does not have anything to do with our repository or software that is already on the .iso then it does not belong in this section. If you are going to install other third party programs then you are on your own. This section is for problems with out software. What really kills me is that for years people have been crying about a package manger in backtrack capable of updates and no we have one and no one wants to use it.

Posting Permissions

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