[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

What's next?



Now that we appear to be at a sustainable 300 MK/sec (which I could
hardly believe at first :-), is there any chance we could be attacking
the 56-bit DES key next?

DES is rumoured to be faster than RC5; also, there are likely to
be optimized assembler versions out there already.

We could also make better use of 64-bit architectures, since DES
uses 64-bit blocks.

However (and this is a very big however), that's still a keyspace of
7.2*10^16 keys to search.  Assuming 10^9 keys/second (three times our
current speed, which may be attainable with a good DES implementation)
that's still a bit over two years; too long by a factor of 10 or so.

Soo....

How fast are current DES implementations?  People could try des.c from
the ssh distribution as a starting point.

How much more computing power could we bring online?

Do people have optimized DES for Alpha, UltraSparc, HP 8000 and all
the other nifty 64-bit architectures?

Does it make sense to use MMX or equivalent for DES?
-- 
Thomas Koenig, [email protected], [email protected].
The joy of engineering is to find a straight line on a double
logarithmic diagram.