Re: Shannon -- the stream cipher, that is



Paul Rubin <http://phr.cx@xxxxxxxxxxxxxx> writes:
"xmath" <xmath.news@xxxxxxxxx> writes:
Salsa20 is quite fast, and has no key/IV setup overhead whatsoever.

True, especially with 8 rounds. But I was hoping for something that
beats RC4.

It's particularly fast on PowerPC G4, where you get less than 280
cycles per 64-byte block, which makes for 4.5 cycles/byte.

Yeah, but I think those are becoming extinct. I guess they still
exist in some embedded products, but Macintoshes don't use them any
more.

Freescale's Power Architecture kit (in particular G4-family, though
usually with limited or no FPU) is exceptionally common in network
infrastructure equipment. There is often some crypto taking place
in those processors. (I believe some have Kasimi, or whatever it's
called, accelerators, for example.)

Phil
--
"Home taping is killing big business profits. We left this side blank
so you can help." -- Dead Kennedys, written upon the B-side of tapes of
/In God We Trust, Inc./.
.



Relevant Pages

  • Re: Shannon -- the stream cipher, that is
    ... beats RC4. ... cycles per 64-byte block, which makes for 4.5 cycles/byte. ... Yeah, but I think those are becoming extinct. ...
    (sci.crypt)
  • Re: 64-bit AES
    ... AES128 206 cycles (12.9 cycles/byte) ... the code gets a bit ugly since some instructions end up in odd places so ...
    (sci.crypt)
  • PolyMAC-1305 [was Re: Pelican MAC [speed and vectors]]
    ... multiplier. ... On the k7 a multiplication takes 6 cycles which is 150 ... cycles/byte on an AMD64. ... Pelican implementation in PORTABLE C gets. ...
    (sci.crypt)
  • Re: doubts about rc4
    ... >for rc4. ... the number of cycles are varying as function ... >realized the array has to be filled with the key size and, in turn, ... Greg Rose ...
    (sci.crypt)
  • Re: AAA, etc.
    ... and you claimed it was M$ fault? ... Yeah it was this one ... ... code exceed cache lines or cross pages. ... a maximal deviation of a few cycles. ...
    (alt.lang.asm)