hashcat Forum

Hashcat have this option. It's useful. Oclhashcat don't have it.
Any chance or interest in being it implemented to oclhashcat as well ?

there would be no point in implementing it in oclHashcat. nor is it really feasible since the rule engine is on the GPU, not CPU. it would be much slower than just using --debug with cpu hashcat.

For old architecture, epixoip is right, but with the refactorization made on v1.02 it might be possible with some effort. Just not sure if its worth it

I would be in favor of it. You can "push" a lot more of rules onto gpu than on CPU. Again, this is all in the goal of determining what would be the best rules to use with each wordlists. Being able to get both the cracked password and the rule that cracked it would be amazing.

High speed would not be an important factor as the final result, the successful list rules, would be
very valuable. Then it would be possible to merge, or not, those successful rules with new ones found
in the same way. Even generated rules, which can go up to xxx k, or whatever limit that can be tested,
can be then identified this way and added to a kind of personal, if not public, "premium" rules file.
This can be stunning.

done in latest oclHashcat beta

FULL SPEED!!

Amazing. I will try it when I will go back home.

Outstanding work atom ! Thank you for the effort to implement this.

I have found a bug with the debug feature (mode 3) in OclHashcat. If you crack a password with 14 or 15 rules you get long lines full of nulls and weird characters. Those cannot be pasted here so they are included in an attached file.[/code]