oclHashcat-plus-0.10-rc3-12.8

[M@LIK]

The lots-of-wordlists bug is fixed in b2

Confirmed. Thanks!

[proinside]

I'm still runing this "old" version, 0.10-rc-3-12.8 and I'm getting a situation
that would like to report to know if this is normal to happen.
We run in command line a normal, no rules, attack.
We get results normally and when we repeat the same command line, which
means using the same dictionaires, the same hashes file (now with less hashes
cause --remove option is used, obviously), if all works fine we get zero found
passwords. That's ok.
Now, we keep working the same way, with the same hashes file, after "cleaning"
with the first simple run attack, but here we only add the best64 rule.
At first we get some results.
We run a second time and we get more results again, not the sames got from the
same run. And this can happen again and again.
This means that the best64 rule isn't full checked against all the hashes (less
than 1,5 millions md5), or I'm doing something wrong ? Can this be considered
normal (even adding the found passwords to a dictionaire), we still get some
results while running again the same simple command line with the --remove
and best64 rule, and aren't the same passwords found in the prevoious run.
I thought, and was expecting that, even after running twice the same command
line with the best 64 rule, no more results would be found. All the possible
passwords would be cracked after the first, or second run, with that rule.
This reminds me that hashcat doesn't remove duplicated hashes, but that's not the
case here.
The best way I see for a consistent result is with no rules.
Moving to a recent version means that 12.8 should be (again, we know, amd)
to do a very clean uninstall, and install 13.1, but before starting with a new
version, I'm curious why this is happening.
Didn't try to disable markov option... can this be the reason ?!

[proinside]

Still using the same dictionaires and remaining hashes file, started
another run, this time with --markov-disable and oclhashcat is yet
founding results, on the second run with this option disable.
So this isn't the cause.
--gpu-loops isn't specified in the command line, only -n option is
set to 8, since the first run, just to specify what is being used,
run after run, always with the same command line.