Date: Sun, 1 May 2011 05:38:48 +0200 From: Łukasz Odzioba <lukas.odzioba@...il.com> To: john-dev@...ts.openwall.com Subject: PHPASS CUDA patch revision 0 Last week i have developed PHPASS patch for JtR. I need to find better way to compare it with oclhashcat, for now it is only JtR benchmark. This is very test-password sensitive ( results may differ by taking another test cases ). I'am getting currently ~27k c/s on 9800GT. It would be nice to see benchmarks result for other nVidia cards so here is test-script for anyone who would want to share results: #!/bin/bash john_target="linux-x86-mmx" john_src="john-1.7.7.tar.gz" john_dir="john-1.7.7" john_path="john-1.7.7-phpasscuda-0.diff" if [ ! -e $john_src ]; then wget http://www.openwall.com/john/g/john-1.7.7.tar.gz fi if [ ! -e $john_dir ]; then tar xvzf john-1.7.7.tar.gz fi if [ ! -e $john_path ]; then wget http://ukasz.sphere.pl/gsoc2011/john-1.7.7-phpasscuda-0.diff fi if [ ! -e ispatched ]; then cd $john_dir patch -p1 -Z < ../john-1.7.7-phpasscuda-0.diff cd .. touch ispatched fi if [ -e ispatched ]; then cd $john_dir/src make clean make $john_target ../run/john -test --format=phpasscuda fi On 64bit systems please change target to linux-x86-64. Running benchmark freezes my system on a two or three seconds, I don't know why. There are few known "problems" with this revision: Number of keys per crypt is constant, what is problematic for grater salt loopCounters (process is beeing killed) $H$ prefixes isn't supported Patch does not check for specific situations as "0*", wrong loopCounter etc. Next revision will solve all of them. Lukas
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.