Left mask + dict usually takes way longer than dict + mask. Not strange, at least
to my use.
When you run Mask + Dict it says "Exhausted" after 1 second doesn't look too
much strange cause you're only using 2 digits and that is done at "speed light",
even by a 5870 (well, 1 second probably is too fast, but still, I do it and it's
very very fast simply adding digits to a dict file).
Try to increment the number of digits on the right mask and see if that second
time remains the same or not. When you start adding digits, you will get longer
run times.
(06-25-2012, 11:17 PM)proinside Wrote: [ -> ]Left mask + dict usually takes way longer than dict + mask. Not strange, at least to my use.
Thanks proinside, I don't understand why what I have quoted above would be the case. Surly there will be the same number of passwords to check regardless of which end you are attaching numbers to ?
Anyway, your suggestion for me to change the mask has revealed the problem !!
OK
If I choose dict - mask with ?d?d it runs for a few seconds.
If I choose mask - dict with ?d?d it runs for 4 minutes.
However ....
If I choose dict - mask with ?d?d?d?d it runs for 12 minutes.
If I choose mask - dict with ?d?d?d?d it runs for 12 minutes.
How strange is that ?
The first batch the dict - mask runs considerably quicker than mask - dict even with the same mask.
The second batch is exactly what I would expect to happen both take roughly the same time.
Thanks for your interest in this proinside.
Does this help atom ?
Yeah please read that post linked from malik, its essential.
However the original bug this thread is about is something different. I think it has to do that in WPA we reject all words if the are < length 8, since this a some WPA restriction. It might be that oclHashcat-plus takes only the length of the hybrid mask and then to decide to reject them all since they are all of the same length, and all < length 8. I did not check it, you should try it. Take ?d?d?d?d?d?d?d?d as Mask. If its not happening then this is the problem.
(06-26-2012, 11:47 AM)atom Wrote: [ -> ]Take ?d?d?d?d?d?d?d?d as Mask. If its not happening then this is the problem.
This is it !! Well done atom.
If I use ?d?d I get the error reported.
If I use ?d?d?d?d?d?d?d?d it works, going to take 57 years but it is working !
My test list for this has words of all lengths. So as you say it must only be reading the mask not the complete password candidate.
Hope this helps and thank you for making me feel less paranoid, I hate it when it is only me with bug reports that don't seem to be reproducible.
Ok, the v0.09 is not in the beta section anymore but in v 0.08 I get the same problem as Hash-it. Mode 7 don't seem to work with heavily iterated algorithm like Wordpress, phpbb3, MD5(apr), etc. It works fine in unsalted hashes like MD5, SHA1 and also in algo like DCC and SMF. Hopes that it helps nailing down the problem.
Those test where done on Windows 7 - 64 bits, Catalyst 12.6 with HD5750 GPU.