I got this report today. The Intel C++ v11 compiler gives better code than Doug's assembler.
From: "Danilo Gligoroski" <firstname.lastname@example.org>
Date: December 6, 2008 1:03:50 AM PST
Subject: Skein is faster than 6.1 cpb with the Intel C++ v11 compiler
Today I was playing with your optimized Skein C version on the NIST reference hardware platform but with the newest Intel C++ v11.
I got 5.99 cpb. Probably by finding different compiler swithches and some other fine tunings the speed can be even more improved.