15 chars limitation
#41
Sry to bring this up again. But nobody has mentioned salts yet. I wouldn't really care about 15+ passwords, but when it comes to salts with lengths of 10+ even 5+ char passwords can't be cracked anymore with oclhashcat. And salted passwords are really hard to crack with cpu haschat, because of the time consumption. As a hash and password-collector i have multiple millions of different salted passwordhashes here that i'm not able to crack because either cpuhashcat is too slow, lacks of dic combination for dynamic pre- & appendsalts and no hybrid attacks, or oclhashcat simply can't crack them because of the length.

You should also consider releasing two different versions of oclhashcat, one that still supports the faster 15char method for better precalculating and another slower version for longer passwords.
I don't know if you are already doing this, but including fix/static salts in the precalculation of the first few cycles would also be great, or are the pre-calculation steps only worth it if the dword blocks for precalculation are padded zeros?

Cheers Wink
#42
(05-22-2013, 07:05 PM)5andr0 Wrote: but when it comes to salts with lengths of 10+ even 5+ char passwords can't be cracked anymore with oclhashcat.
Hmm, why they couldn't be ?

Well, this seems to be a really serious question. We need longer passwords, but we need the best speeds as well. Probably, we can't get both.

The thing I would do if I were on atom's place is I would introduce new option (--max-length or something like that...) which would compile new kernel with maximal number of rounds what are necessary for computing of password of maximal length. This option would be defaultly set to maximal supported length.

But I bet atom'll come up with something brilliant.
#43
Smile 
I keep begging for the end of this char limitation, at least for WPA. :'-(
I have several dictionaries waiting for it. For example: here in my country every network named "WLAN_XXXX" starting with BSSID 00:19:15 has a key included in a well known 200MB (10 millions) dictionary of passwords... each one of them 20 chars long. :-(

Just a "promenade" for OCLHashCat-Plus... when 15 chars length will be erradicated.
#44
support for length > 15 is my primary goal for next release
#45
atom, will you do something as I mentioned in my last reply ? Please, tell us, I'm curious Big Grin
#46
Thumbs Up 
(05-27-2013, 06:21 PM)Kuci Wrote: atom, will you do something as I mentioned in my last reply ? Please, tell us, I'm curious Big Grin

Give Atom a breath, Kuci. :-)
He could have too much work load to come chatting with us. I do prefer he will invest the time on the next release, that we all wait with hope!
Anyway we will see all these data on "changes.txt".

Go on, Atom!!! The forum is with you! :-D
#47
Thumbs Up 
@atom if you ever get >15 char limit support for WPA in for hashcat's next release I wouldn't mind tipping you with some bitcoins Smile