ryanlim is amazed that people still look up my John the Ripper (MPI parallelized) code.

19 responses to this plurk.

• 
  August 13, 2008 at 18:53 martinbogo Are you kidding? John the Ripper ROCKS when run w/ MPI
• 
  August 13, 2008 at 18:54 martinbogo Scyld ( Or Clusterware ) + MPI) with myrinet or infiniband ... and a couple hundred machines?
• 
  August 13, 2008 at 18:55 martinbogo ... Mmmmm..
• 
  August 13, 2008 at 18:55 ryanlim says i dunno what people's setups are
• 
  August 13, 2008 at 18:57 ClubEddy says you are the author of John the Ripper??
• 
  August 13, 2008 at 18:57 martinbogo well, here's an example : A cluster running at an Agency That Shall Not Be Named:
• 
  August 13, 2008 at 18:56 martinbogo 244 quad socket dual core 64 bit systems
• 
  August 13, 2008 at 18:58 martinbogo running 2.4Ghz, on infiniband (using mvapich) on Scyld. just imagine.
• 
  August 13, 2008 at 18:58 ryanlim says ClubEddy: no. but i made parallelized version
• 
  August 13, 2008 at 18:59 ryanlim says yea. i can imagine
• 
  August 13, 2008 at 18:59 ryanlim says the performance should scale up almost linearly with the number of procs you throw at it.
• 
  August 13, 2008 at 19:00 martinbogo the performance is measured not in kilochars, but in gigachars/sec
• 
  August 13, 2008 at 19:00 ryanlim says this is what i wrote many years ago, don't laugh www.ryanlim.com/personal/jtr-mpi/report.pdf
• 
  August 13, 2008 at 19:05 martinbogo That's a wonderful paper!
• 
  August 13, 2008 at 19:05 martinbogo Short, to the point .. clearly written...
• 
  August 13, 2008 at 19:26 calebcassel says I've used John the Ripper before! That was back in 1997 though so I probably wasn't using yours
• 
  August 13, 2008 at 19:30 Telperion says ooh...smells like cryptanalysis
• 
  August 15, 2008 at 05:45 conversationing says I'm pretty sure my friend used JtR-MPI for his honours thesis (scissec.scis.ecu.edu.au/conference_proceedings/2004/forensics/Frichot-1.pdf)
• 
  August 15, 2008 at 11:31 ryanlim says conversationing: i see my name!