Author: Andrew Shapira
Date: 23:40:22 09/21/05
Go up one level in this thread
Yeah, I know about http://graphics.stanford.edu/~seander/bithacks.html - I have used it and some other web sites extensively. I have racked my brains for 2 days, on the couch, in the shower, in bed, while walking around, and still haven't figured it out. My sense is that it seems doable, and that a solution could involve subtracting x from (2^n)-1. This might be a high-order analog of the x-1 term that occurs in the expression x^(x-1) that comes up in low-order bit computations. That sucks about that patent! On September 22, 2005 at 01:11:06, Chan Rasjid wrote: >On September 21, 2005 at 23:31:56, Andrew Shapira wrote: >>What you ask is what all the chess programmers using bitboards need badly. >I think the possibility is slim as none in CCC ever mentioned any success so >Gerd had to get alternative. > >http://graphics.stanford.edu/~seander/bithacks.html >This link had all the things about log base 2 etc.. but I think 3- 7 operations >for 1 byte w/o conditionsals. > >BTW if you are in the US you cannot use :- >#define ABS_INTEGER(x) ((+1|(x>>SIZE_OF_INT_LESS1))*x) >A Russian got the US patent. > > >Rasjid
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.