Niklas Mehner wrote:
> CPU TIME (ms) BEGIN (total = 431) Thu Jan 10 16:32:28 2002
> rank self accum count trace method
> 3 7,19% 83,76% 2822912 2169 Freenet.crypt.ciphers.Twofish_Algorithm._b
> Since I did not want to optimize wait() or sleep() I looked at _b() a
> small patch.
> This improves the _b() performance by about 25% (should be about 5-10%
> for the whole startup process).
> private static final int _b( int x, int N) {
> - int result = 0;
> - switch (N%4) {
> - case 0: result = b0(x); break;
> - case 1: result = b1(x); break;
> - case 2: result = b2(x); break;
> - case 3: result = b3(x); break;
> - }
> - return result;
> + return (x >>> ((N%4) * 8)) & 0xFF;
This patch makes sense to me (pretty straightforward) and if the savings
are really that large, I propose that somebody checks it into the
new_datastore and HEAD branches of CVS. I am al for it...
Any volunteers?
Sebastian
_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl
