LeeBailey
07-03-2006, 04:15 PM
Nuts.
I've spent the last few days researching rainbow tables, and computing relative size-vs-speed benefits, not to mention torrenting MatLab, and I've come to the following conclusion:
To precompute MD5 hashes with 98.6% accuracy from 0:7 characters for upper/lower alpha-numeric (56 keys, keyspace 3579345993195) would take 100gb of space and 183 days of computation on my computer (split across 3 computers, one with two cores, all rated around 3300 bogomips, that's 45.75 days).
Really, I can't expect the campers to care much, but Nick, this one's for you, buddy.
(On the other hand, 0:7 LM hashes, which are only 36 keys due to LM's auto-uppercasing, and 98.8% success, would only take 4.2 days, or 1 day per processor, and 2gb of space. I think we have a winner!)
I've spent the last few days researching rainbow tables, and computing relative size-vs-speed benefits, not to mention torrenting MatLab, and I've come to the following conclusion:
To precompute MD5 hashes with 98.6% accuracy from 0:7 characters for upper/lower alpha-numeric (56 keys, keyspace 3579345993195) would take 100gb of space and 183 days of computation on my computer (split across 3 computers, one with two cores, all rated around 3300 bogomips, that's 45.75 days).
Really, I can't expect the campers to care much, but Nick, this one's for you, buddy.
(On the other hand, 0:7 LM hashes, which are only 36 keys due to LM's auto-uppercasing, and 98.8% success, would only take 4.2 days, or 1 day per processor, and 2gb of space. I think we have a winner!)