There are not only ways of optimization via C, but it's also possible to optimize assembly using automated tools. While the use of such tools is limited, it may prove quite handy, i leave a link to an adjunct website to the book Hacker's Delight where one can find "A Hacker's Assistant" a small tool that uses brute force to find equivalent, more eficient assembly code in RISC processors, obviously not all CPUs have the same instruction set, but since inner loop instructions of crutial speed areas are usually simple but costly in time, this tool might reduce the number of instructions and, therefore, clocks taken.
Please note, and i cite, "There is nothing here about hacking in the sense of breaking into computers." in the website i provide. Also note, reading the book will help understanding what you're doing XD (but if you are familiar with SWAR you may find it has information you knew already)
Sometimes tool yields some interesting and unexpected (but quite powerful) results, have fun.
PS: Oh.. like i said, brute force algorithms are used... so don't expect to optimize large amounts of code...
--Zebarnabe 15:48, October 7, 2009 (UTC) zebarnabe