Author: Gerd Isenberg
Date: 22:16:38 02/28/06
Go up one level in this thread
On February 28, 2006 at 17:51:34, Randall Shane wrote:
>Gerd,
>
>Since the DeBruijn approach looks interesting to me, I thought I'd give it a
>try.
>
>I tried running the code given 2 messages above for the optimized De Bruijn
>generator, with the following derived class and main function :
>
>
>class DBG : public DeBruijnGenerator
>{
>public:
> DBG() {}
>
> void deBruijnFound(BitBoard deBruijn) const
> {
> printf("%-9d : %016I64X\n", count+1, deBruijn);
> // Running with the Microsoft compiler on a Pentium IV machine
> }
>};
>
>void main(int argc, char* argv[])
>{
> DBG *dbg = new DBG();
> dbg->genDeBruijn(3);
> fflush(stdout);
>}
>
>just to check it out. I have it generating De Bruijn sequences of length 8 (all
>patterns of 3 bits).
>
>I get the results :
>
>1 : 0000000000000017
>2 : 000000000000003D
>
>The second one appears to be in error -- it should be hex 1D (00011101).
>
>I'd try one of the earlier generators, but I can't find them :-(.
>
>Anyway, though you'd like to know..
Thanks - may be introduced by the depth-2 optimization...
Indeed with some additional cost the if "(depth>2)" case may be skipped.
Will have a closer look...
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.