Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <55BB1F3C.9060206@gmail.com>
Date: Fri, 31 Jul 2015 09:09:48 +0200
From: Marek Wrzosek <marek.wrzosek@...il.com>
To: john-users@...ts.openwall.com
Subject: Re: John the ripper with Tesla GPUs on Debian

W dniu 30.07.2015 o 23:28, Viktor Gazdag pisze:
> Hi all,
> 
> 
> I've installed the latest Tesla driver from nvidia site, not the
> nvidia vga driver and not from repo.
> We downloaded and installed the ubuntu 14.04 deb package and the
> dependences with aptitude and some dependencies with ubuntu packages,
> because some version weren't high enough or weren't available. Maybe
> the cuda*.run installer would be better, but I haven't had that much
> free space left to try to install it.
> We've installed librexgen from github, but john didn't find one lib,
> because it was in an another directory (I think it was in the
> librexgen/c/ directory).
> 
> Before run configure, I've added the following env (these are needed
> for running john, too):
> export CC=gcc-4.9
> export PATH=$PATH:/usr/local/cuda-7.0/bin
> export LD_LIBRARY_PATH=:/usr/local/cuda-7.0/lib64
> export PATH=$PATH:/usr/local/cuda-7.0/targets/x86_64-linux/include
> export LD_LIBRARY_PATH=:/usr/local/cuda-7.0/targets/x86_64-linux/lib
> export PATH=$PATH:/usr/lib/gcc/x86_64-linux-gnu/4.9
> 
> The last export is needed, because system didn't find cc1plus to compile things.
> 
> ./configure --enable-experimental-code
> make clean && make -s
> 
> After it, I could list cuda devices:
> 
> ./john --list=cuda-devices
> CUDA runtime version 7.0
> CUDA driver version 7.0
> 8 CUDA devices found:
> 
> CUDA Device #0
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           0d:00.0
>     NVML id:                       2
>     Fan speed:                     n/a
>     GPU temp:                      24°C
>     Utilization:                   0%
> 
> 
> CUDA Device #1
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           0e:00.0
>     NVML id:                       3
>     Fan speed:                     n/a
>     GPU temp:                      24°C
>     Utilization:                   0%
> 
> 
> CUDA Device #2
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           09:00.0
>     NVML id:                       0
>     Fan speed:                     n/a
>     GPU temp:                      23°C
>     Utilization:                   0%
> 
> 
> CUDA Device #3
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           0a:00.0
>     NVML id:                       1
>     Fan speed:                     n/a
>     GPU temp:                      22°C
>     Utilization:                   0%
> 
> 
> CUDA Device #4
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           28:00.0
>     NVML id:                       4
>     Fan speed:                     n/a
>     GPU temp:                      27°C
>     Utilization:                   0%
> 
> 
> CUDA Device #5
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           2b:00.0
>     NVML id:                       5
>     Fan speed:                     n/a
>     GPU temp:                      27°C
>     Utilization:                   0%
> 
> 
> CUDA Device #6
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           30:00.0
>     NVML id:                       6
>     Fan speed:                     n/a
>     GPU temp:                      27°C
>     Utilization:                   0%
> 
> 
> CUDA Device #7
>     Name:                          Tesla K20Xm
>     Type:                          discrete
>     Compute capability:            3.5 (sm_35)
>     Number of stream processors:   2688 (14 x 192)
>     Clock rate:                    732 Mhz
>     Memory clock rate (peak)       2600 Mhz
>     Memory bus width               384 bits
>     Peak memory bandwidth:         249 GB/s
>     Total global memory:           5.0 GB (ECC)
>     Total shared memory per block: 48.0 KB
>     Total constant memory:         64.0 KB
>     L2 cache size                  1.0 MB
>     Kernel execution timeout:      No
>     Concurrent copy and execution: Bi-directional
>     Concurrent kernels support:    Yes
>     Warp size:                     32
>     Max. GPRs/thread block         65536
>     Max. threads per block         1024
>     Max. resident threads per MP   2048
>     PCI device topology:           33:00.0
>     NVML id:                       7
>     Fan speed:                     n/a
>     GPU temp:                      26°C
>     Utilization:                   4%
> 
> Best regards
> woodspeed
> 
Hi

About librexgen... What library was missing? As far, as I know,
bleeding-jumbo requires exactly 1.2.5 version of librexgen now, but this
is not mandatory and there are some problems with it.
Installing this version of librexgen is easy - just run install.sh
(without sudo but it's using sudo to run make install, so for debian
you'll need to run make install using su instead of sudo). Install
script (using make install command) will copy header files to
/usr/local/include/librexgen and librexgen.so.1.2.5 to /usr/local/lib
(and will create symlinks librexgen.so.1.2 and librexgen.so) but
ldconfig you'll need to run by hand. Then configure script of john
should find librexgen without problems.

Best Regards
-- 
Marek Wrzosek
marek.wrzosek@...il.com

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.