Re: [md5] string giving specific hash



rossum <rossum48@xxxxxxxxxxxx> writes:

On 23 Sep 2006 07:22:13 -0700, bili.pl@xxxxxxxxx wrote:

Does anybody know any string which gives MD5 hash containing only
decimal digits (all 32 digits 0...9)? Or does anybody know how can I
find it using eg. brute-force method?

I found one after 10212 generations of a very simple genetic
algorithm:

String: 8F B2 6C 31 6F F9 35 88 96 FF 20 C4 FD 31 6F F9
8F B2 99 02 88 31 AB 77 36 1E 96 F2 33 E3 3C

Hash: 85 71 09 26 50 22 13 40 25 17 88 89 30 85 18 95

rossum


uh, at this point I am very worried. What do you mean a genetic algorithm?
Could you post it?
If there is a way of getting md5 to output a hash which is not random (
which is what a genetic algorithm sounds like to me) then MD5 is seriously
broken. Ie, there should be nothing in hash A to give any clue as to what
hash B would be no matter what the relation of B to A ( except of course if
B=A). If A and its hash can be used to predict any feature of B (eg select
B so that its probability of all decimal digits is increased), no matter
what B is, then MD5 is sick.

If by "generations" you mean that you ran MD5 10212 times and got an all
numeric hash ( assuming by that you mean that the output of MD5 was the hex
value of the hash) then again, MD5 is sick, since the probability of
getting an all digit hex hash in less than or equal to 20000 trials should
only be about .5% . Ie, the probability that MD5 is random is rejected at
the 99.5% confidence level.

.