Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Fri, 29 May 2015 10:46:05 +0300
From: Solar Designer <solar@...nwall.com>
To: Alain Espinosa <alainesp@...ta.cu>
Cc: john-dev@...ts.openwall.com
Subject: Re: bitslice SHA-256

On Fri, May 29, 2015 at 12:55:18AM -0400, Alain Espinosa wrote:
> The most expensive part in this implementation is the sum (5 instructions) with can be reduced in common architectures. In Neon we can reduce the carries calculation from 3 to 2 instructions (this is basically a MAJ function that can be implemented with a bitselect+xor). In AVX512 we can reduce the sum from 5 to 2 instructions with the generalized 3 operand boolean instruction.

Oh, so you already explored this.  Cool.

What you say about NEON is also applicable to XOP.

Alexander

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ