I just wrote an MD5 hash brute-forcer. It's Linux native, but runs perfectly (and just as fast) under LINE in Windows. It's written in assembly (duh), and there are two binaries in the archive: md5crack, and md5crackl.

md5crack generates alpha-numeric characters to crack. This can be used if you don't know the type of characters in the message.

md5crackl generates only lower-case letters, and also numbers. This is how Turok's MD5 cracker does it. Doin it this way is exponentially faster at cracking hashes, when the string has no upper-case letters in it.

I ran the same hashes on this, against the one that Turok wrote (thanks for the idea), and it's _much_ faster.

Cracking 7b2fa6be6b14cb83c15e78a5e9c8fa5b with Turok's cracker took about 10 seconds on my machine... using my cracker (md5crackl), however, took only 4 seconds. Cracking a hash made from "hello" took 6 seconds on his program, and 3 seconds on mine.

Hopefully this program can be of use to you all. I hope you like it. Comments/suggestions, please tell. Included in the archive is a README.

... and where is sources ??

Oh its so slow



---------------------------------------
running md5 hash brute-forcer...

MD5 hash successfully cracked.
Message: wasmru

Time elapsed: 354.410s
----------------------------------------

Runned on a Opteron 3,6 Dual

sylwek32: Of course it's slow, it uses pure brute-force. So it needs to try about (26+26+10)^6 combinations, maximum, on that 6-char string.

Did you use md5crack, or md5crackl? I can crack a hash made from "howdym" in 112s using md5crackl on a 3.2GHz P4, and in 214s using Turoks cracker.

How long did it take you to crack that same hash using Turok's? It should be about 2x slower... Just curious.

Just a small bug fixed... Just download the file in my first post again.

I'm not sure what is the actual difference in speed of the programs becose your program first tests a-z and my program first tests 0-9 so a different password should be required to test each of both programs. Let say you can test my tool for hash of string "000000" and test your tool for hash of string "aaaaaa". I think that should give both tools the same test of dificulty.

Ahh, I did not know that. You are right then... cracking a hash made from "000000" on your cracker takes me 8.9 seconds. Cracking a hash made from "aaaaaa" takes 13.4 seconds on my cracker.

However, a hash made from "010101" takes 23 seconds on your cracker, and a hash made from "ababab" takes 13.8 seconds on my cracker.

I think that is becouse i use left-to-right aproach and you use right-to-left aproach so in fact if u want to compare mine "010101" you have to compare it with yours "bababa" and if you want to compare yours "ababab" you have to compare it with mine "101010"

Hey Patrick, that is indeed a powerful md5 cracker, I was wondering if you mind releasing the source for it, im sure a lot of people could learn from it. (I'm sure I could )