Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: R. Hyatt, Crafty style move generation

Author: Larry Griffiths

Date: 16:08:55 01/23/01

Go up one level in this thread


On January 23, 2001 at 16:37:23, Dann Corbit wrote:

>On January 23, 2001 at 16:11:12, Larry Griffiths wrote:
>[snip]
>>I would think that increasing the performance of my move generation would help
>>the search since making and unmaking moves is only a subset of the moves that
>>are generated.  This increase in Moves-Per-Second is mainly due to combining all
>>the generate code from my Piece classes into ONE big chunk of code.  I think
>>that I am getting a performance increase due to the elimination of calling
>>Genmoves for each piece.
>
>Lots of compilers have a setting that allows the compiler to inline like a
>madman.  I prefer that approach, since it makes the code a bit easier to
>understand.

Derived classes that contain different versions of a method do not fit into the
inline picture.  I have a derived class for each piece and each GenMove method
is different.  The call to the class looks the same though I.E.

   Piece->GenerateMoves();

The Borland CBuilder product does not let me force inline, and does not allow
inline assembly in templates or inline functions so I have to work around these
restrictions.

Larry.





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.