john-users - Re: MultiGPU support in Magnum?

Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d3406e5b246dd440334c7bb4e1d9454d@smtp.hushmail.com>
Date: Thu, 19 May 2016 22:27:08 +0200
From: magnum <john.magnum@...hmail.com>
To: john-users@...ts.openwall.com
Subject: Re: MultiGPU support in Magnum?

On 2016-05-19 11:13, Doc Case wrote:
> Thank you Magnum.
>
> It was pretty straight-forward, my fault not to check on docs... Anyhow,
>  something is still not right. This is the output I get, seems that a
>  seg-fault happen on dev0, and fall back using just a single GPU...

I haven't seen anything like that. There's no reason this 
multidevice-using-fork thing should cause problems for a system which 
works fine using one GPU at a time. Does this happen with other formats 
(than DMG) as well?

magnum


> ./john -fork=2 -dev=gpu,acc -format=dmg-opencl ~/Downloads/hash.txt
>
> Using default input encoding: UTF-8
>
> Loaded 1 password hash (dmg-opencl, Apple DMG [PBKDF2-SHA1 OpenCL 3DES/AES])
>
> Will run 4 OpenMP threads per process (8 total across 2 processes)
>
> Node numbers 1-2 of 2 (fork)
>
> Device 0: GeForce GTX 970
>
> [Tappo:03548] *** Process received signal ***
>
> [Tappo:03548] Signal: Segmentation fault (11)
>
> [Tappo:03548] Signal code:  (128)
>
> [Tappo:03548] Failing at address: (nil)
>
> [Tappo:03548] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x113d0)[0x7fab600943d0]
>
> [Tappo:03548] [ 1] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x96e9dc)[0x7fab547169dc]
>
> [Tappo:03548] [ 2] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x96ee95)[0x7fab54716e95]
>
> [Tappo:03548] [ 3] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x96f0e5)[0x7fab547170e5]
>
> [Tappo:03548] [ 4] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x507e30)[0x7fab542afe30]
>
> [Tappo:03548] [ 5] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x5086e2)[0x7fab542b06e2]
>
> [Tappo:03548] [ 6] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x512f6a)[0x7fab542baf6a]
>
> [Tappo:03548] [ 7] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x51338a)[0x7fab542bb38a]
>
> [Tappo:03548] [ 8] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x97ae45)[0x7fab54722e45]
>
> [Tappo:03548] [ 9] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x97b578)[0x7fab54723578]
>
> [Tappo:03548] [10] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x97bcbe)[0x7fab54723cbe]
>
> [Tappo:03548] [11] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x3f88a9)[0x7fab541a08a9]
>
> [Tappo:03548] [12] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x3f9429)[0x7fab541a1429]
>
> [Tappo:03548] [13] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x16c355)[0x7fab53f14355]
>
> [Tappo:03548] [14] /usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1(+0x16b47d)[0x7fab53f1347d]
>
> [Tappo:03548] [15] ./john[0x73c04c]
>
> [Tappo:03548] [16] ./john[0x73da8c]
>
> [Tappo:03548] [17] ./john[0x6ea10a]
>
> [Tappo:03548] [18] ./john[0x6f2355]
>
> [Tappo:03548] [19] ./john[0x6f3f78]
>
> [Tappo:03548] [20] ./john[0x6f42e6]
>
> [Tappo:03548] [21] ./john[0x6f5239]
>
> [Tappo:03548] [22] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf0)[0x7fab5fcda830]
>
> [Tappo:03548] [23] ./john[0x407c39]
>
> [Tappo:03548] *** End of error message ***
>
>
>
>
> Sent using Zoho Mail
>
>
>
>
>
>
> ---- On Thu, 19 May 2016 10:37:25 +0200 magnum &lt;john.magnum@...hmail.com&gt;wrote ----
>
>
>
>
> On 2016-05-19 02:06, Doc Case wrote:
>
> &gt; Hi to everyone,
>
> &gt; I think i'm missing something... I have 2 GPUs in my system, a couple of GeForce GTX970 and I've built from bleeding john with support for openCL, cuda and OpenMPI.
>
> &gt; I'm tryng to crack a SHA1 3DES/AES, but john works only with one GPU.
>
> &gt; In nvidia-settings i can see the temperature rising just on one of the two cards... any suggestion on what i'm missing? I need to use two different session, one for each card? if so, can you point some instructions to do so?
>
>
>
> doc/README-OPENCL, section "Multi-device support".
>
>
>
> magnum
>
>
>
>
>
>
>
>
>
Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.