Cryptohaze.com

Yes DES-based Unix crypt() is like 30+ yrs old. However, there are still many large legacy UNIX systems out there, and it is still the default in Solaris 10. Also dont forget about the recent security breakin at Twitter, which shows that many companies still dont enforce Strong password policy for all employees. So even with salts, large brute force (possibly rainbow table assisted) dictionary attacks on DES should still be quite feasible, esp with the gp-gpu approaches that this forum centers around.

Current DES stats:
Quad-core Q6700, ~8.8 Mk/s, (using bitslicing)
Playstation 3, ~11.5 Mk/s, (using bitslicing)
DESCHALL Project, ~7 Gk/s, worse case ~120 days, idle time of clients
distributed.net, ~28 Gk/s, worse case ~40 days, idle time of clients
EFF Deepcrack, ~93 Gk/s, worst case ~9 days, cost $250,000 cust hardware
COPACOBANA, ~64 Gk/s, worst case ~13 days, cost $10,000, FPG based

I for one would like to see how well an optimized GPU could do at this...

It's certainly something I intend to explore.

Do you have source for DES based crypt laying around?

Or any of the bitsliced code?

I believe you can find that in the source for John the Ripper, at http://www.openwall.com/john/

Bitweasil wrote:It's certainly something I intend to explore.

Do you have source for DES based crypt laying around?

Or any of the bitsliced code?

Is there any way to create rainbow table to crack DES or 3DES ? The source code is in OPENSLL source code .