Andrew Warren wrote:
>
> Dudes:
>
> A few days ago, someone on the list mentioned TEA (the "Tiny
> Encryption Algorithm"), developed by Wheeled and Needham at
> Cambridge.

I managed it working on Friday night :-) (I have Internet access at work
so I don't read mail in the week ends).

> I've just thrown together a PIC16 implementation.  The PIC isn't
> really well-suited for the job -- TEA is optimized for 32-bit
> processors with fast barrel-shifters -- but it still manages to
> encrypt or decrypt a 64-bit block in about 6600 cycles (which works
> out to 6K bytes per second at 20 MHz).

I got 6538 cycles for encipher and a little more for decipher (due of
the subtractions instead of additions: they are not commutative!).
I am sure I can shrink the execution time using ad-hoc shift and add
routines instead of the generalized ones I use currently.

> If any of you wants a copy, let me know... I've sent my code to Dr.
> Simon Shepherd for verification; as soon as he agrees that it's a
> correct implementation, I'll be happy to distribute it.

If you like I can send you my code while you send me yours... For a
quick check: enciphering an all zero's data block using an all zero's
key gives a 0xed285da1 0x455b33c1 cipher block.

> Some details:
>
>     The "encode" and "decode" routines occupy 275 words of code
>     space (each occupies about 200 words alone, but they share some
>     large subroutines).

269 words!! (I win!!! :-)))

>     37 bytes of RAM are required:  16 for the 128-bit key, 8 for the
>     64-bit plain/ciphertext, and 13 for local variables.

Oops. I used the same for the key and plaintext/ciphertext but I waste
17 bytes for local use (I know I can save that 32 bit temp register but
I was to tired last Friday...)

> I'm thinking of doing a PIC17 version; it should be significantly
> faster and MUCH shorter... Is anyone interested?
>
> Alternatively, is anyone interested in a PIC16 version that'd be
> WAY slower, but would fit in fewer than 100 words?

Strange. I was thinking of a (maybe) longer but FASTER version.

> -Andy
>
> === Andrew Warren - fastfwd@ix.netcom.com
> === Fast Forward Engineering - Vista, California
> === http://www.geocities.com/SiliconValley/2499

BTW: if there is someone willing to have my code I can post it here (it
is a good practice?) or I can send individual copy.

Ciao
  Marco

----
Marco DI LEO                  email: m.dileo@sistinf.it
Sistemi Informativi S.p.A.    tel:   +39 6 50292 300
V. Elio Vittorini, 129        fax:   +39 6 5015991
I-00144 Roma
Italy