john-users - issues with mpirun

Openwall wordlists collection for password cracking (20+ languages)
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Thu, 2 May 2013 19:29:35 +0100 (BST)
From: T E <tobias2esc@...oo.com>
To: "john-users@...ts.openwall.com" <john-users@...ts.openwall.com>
Subject: issues with mpirun

This is my first posting to this mailing list. Hi everyone! 


I do have some issues with mpirun to run JohnTheRipper on multiple cores. I have a cpu with 8 cores, and atm I'm mainly interested in cracking raw-md5 passwords.
My system runs an ubuntu 12.04, and I tested the most recent JTR versions 1.7.9 jumbo 5 and jumbo 7. I modified the Makefile in the src-folder in such a way that those lines necessary for mpi and omp are uncommented:

## For experimental MPI_Abort support, add -DJOHN_MPI_ABORT too.
CC = mpicc -DHAVE_MPI
MPIOBJ = john-mpi.o

OMPFLAGS =
# gcc with OpenMP
OMPFLAGS = -fopenmp
OMPFLAGS = -fopenmp -msse2

I compile with the flag "linux-x86-64", and there are no errors at compile-time. 

OMP works fine, as the results of the test-benchmarking indicate. However, as raw-md5 doesn't seem to be supported by OMP by now, I tried to start  john using mpirun:

mpirun -n 4 ./john hash.txt

(mpirun --version
mpirun (Open MPI) 1.4.3

Report bugs to http://www.open-mpi.org/community/help/)


This is described at http://blog.thireus.com/john-the-ripped-steak-and-french-fries-with-salt-and-pepper-sauce-for-hungry-password-crackers

However, in my case it is not possible to start multiple threads:

Loaded 2 password hashes with no different salts (Raw MD5 [128/128 SSE2 intrinsics 12x])
Remaining 1 password hashes with no different salts
Loaded 2 password hashes with no different salts (Raw MD5 [128/128 SSE2 intrinsics 12x])
Loaded 2 password hashes with no different salts (Raw MD5 [128/128 SSE2 intrinsics 12x])
Loaded 2 password hashes with no different salts (Raw MD5 [128/128 SSE2 intrinsics 12x])
Remaining 1 password hashes with no different salts
Remaining 1 password hashes with no different salts
Remaining 1 password hashes with no different salts
Node 0@... Crash recovery file is locked: ./john.rec
Node 0@... Crash recovery file is locked: ./john.rec
Node 0@... Crash recovery file is locked: ./john.rec


Same happens when I start a benchmark-testing:
mpirun -n 4 ./john --test

Traditional DES [128/128 BS SSE2-16]... Benchmarking: Traditional DES [128/128 BS SSE2-16]... Benchmarking: Traditional DES [128mpirun: killing job...nchmarking: Traditional DES [128/128 BS SSE2-16]...

Unfortunately,every testcase starts 4 times.

Does anyone has an idea why I have these issues with mpirun & john?


Thanks in advance!
Tobias