Purge rule?

[Kgx Pnqvhm]

Is the purge rule (@X) ever going to be implemented in the GPU hashcat(s)?

As I mentioned in a different post, I am looking into mining pass phrases from web sources, and would store them with the spaces, to then have rules act on them.

Purging the spaces would be one of the actions, but the GPU hashcat doesn't have that rule, so I would need to store a second copy of the phrase, with the spaces removed, as a workaround.

[mastercracker]

I totally agree. It would a huge hard disk space saver. Just need to know if it is feasible on GPU.

[atom]

It's not possible, at least not for the old vector datatype based GPUs VLIW4, VLIW5, sm_21. That is because the purge rules changes the length of the candidate depending on the candidate value. It would be possible for new GPU types that are almost all scalar datatypes, but it would be to much overhead to maintain two different rule engines.

[mastercracker]

Thanks atom. What do you guys think about this solution: for unsalted hashes, run hashcat with the wordlist and purge rules. This should not be costly and finish in a reasonable time. For salted hashes or heavy iterated hashes, uses hashcat with the purge rules in stdout and use it to pipe in oclhashcat. Would this speed the cracking over hashcat alone?

[atom]

I'm not sure if I understood this fully. However, with CPU based algorithms (slow ones, iterated) you can already use the purge rule since oclHashcat uses hashcat rule engine in that case.

[mastercracker]

I'm not sure if I understood this fully. However, with CPU based algorithms (slow ones, iterated) you can already use the purge rule since oclHashcat uses hashcat rule engine in that case.

Not sure that I understood you correctly. Are you saying that if I use the rule "@ " to purge space in oclhashcat for -m 0, it will be rejected but if I use it on -m 400, it will run?

[atom]

yes, that should work