|
Message-ID: <20170814171738.GA2239@openwall.com> Date: Mon, 14 Aug 2017 19:17:38 +0200 From: Solar Designer <solar@...nwall.com> To: john-users@...ts.openwall.com Subject: Re: JtR job with Tesla K80 Nvidia cards not showing status of all devices On Sun, Aug 13, 2017 at 11:06:58PM +0200, Luis Rocha wrote: > Trying last JtR version from git with couple of Tesla K80 Nvidia cards and > for some unknown reason to me, when I try to check the status of the job , > it only prints the status of 1 device where I think it should print the > status of all the devices, right? Is this a cosmetic issue or something > else? In your case, this is a bug - as far as I can tell from what you posted, JtR was only using 1 GPU, and I've just reproduced what looks like the same problem. (To better debug issues like this in the future, please include the output of "nvidia-smi" or "aticonfig --adapter=all --odgc --odgt", as appropriate, obtained while JtR is running.) Please try this: git revert -n 0168a45a145d3c3ff714dd9f6bf73128f5408bac Does it help? (It did here, and I think magnum is to investigate what's wrong with that commit and fix it. Thanks.) > $ ./john /tmp/hash --format=salted-sha1-opencl --fork=2 > Using default input encoding: UTF-8 > Loaded 1 password hash (salted-sha1-opencl [SHA1 OpenCL]) FWIW, there's no point in using a GPU for a run like this, except for testing. You'll achieve better speeds on CPU. To use a GPU on a fast hash efficiently, you need "--mask" (for formats where we do support on-GPU mask, including this one), perhaps along with another cracking mode (if so, you may refer to the other mode's input as "?w" in the mask) or/and many salts (for hash types that are fast yet salted, like this one). I guess you knew that, but thought I'd point it out anyway. > 1 0g 0:00:01:42 3/3 0g/s 38119Kp/s 38119Kc/s 38119KC/s rummbi7..rtbasaw These speeds are on par with one CPU core (not even one CPU). Alexander
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.