Author: Chan Rasjid
Date: 22:11:06 09/21/05
Go up one level in this thread
On September 21, 2005 at 23:31:56, Andrew Shapira wrote: >Hi. I'm interested mainly in ideas - the details of these things can be worked >out later. So being off by 1 somewhere does not really concern me right now. I >have not heard anything about what I asked about - which is, how, in O(1) time, >do you find the mask of the highest bit using only C's + - ^ ~ << >> operators? >I thought this was clear in my original post, but maybe somehow it wasn't. Or >maybe you didn't read it. The whole point is to do it quickly and portably. If >I wanted to do it in an unportable way I would use machine instructions for >finding the highest significant bit, or something similar. If I wanted to do it >relatively slowly, say in O(lg N) time and not O(1) time, I would do it using >the >>1, >>2, >>4 method. It is of some interest to do this using * or / >although that is not the main question here. And I am definitely not interested >at the moment in things that will only work on one machine. 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.