Has anyone tried implementing DES/3DES in assembly? I need one, and was going to do it myself, but just thought it would be good to check if someone got it already.

boss says DES is not safe anymore... and that you should use AES

Not that I have choice.
It's for SmartCards, BTW.

PS. As for AES/Rijandel, we've got it already on the board somewhere...

ummm, some HW-close smartcard driver coding? mmmmmmm....

Tomasz, try looking for "john the ripper", a unix password cracker... it has some DES routines that should be very optimized, or so it's said.

There are lots of crypto implementations on Witeg's site. All are in masm, but written very, very well and thus easy to make fasm-compilable. There's no DES/3DES, but I found:

I found some here, but it's GAS syntax, which is strange enough that I finally done it myself.
It's far from being optimized but at least it appears to be working, and that's enough for me right now.

I'm attaching it here in case of someone else searching for such solution. Note: the provided functions encode/decode just 64-bit blocks given in EDX:EAX. But it should easy to put them in a loop (combined with XOR operation for CBC) to make larger encodings.

When you get the need for speed, google for "DES bitslice", and perhaps add MMX or DES to the party too

try this DES, it's masm syntax

...and fasm version: http://www.wasm.ru/src/9/des.rar

check this ftp's for crypto-information and ready algorithms in c++/c:
ftp.funet.fi
ftp.zedz.net

and another