Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: overhead of

Author: Brian McKinley

Date: 16:05:25 01/25/98

Go up one level in this thread



On January 25, 1998 at 17:55:45, James Long wrote:

>A couple months ago I decided to completely rewrite
>my chess program "Tristram."  One of the things I'm
>doing is moving from C to C++.
>
>Now I'm ready to write some move generation routines.
>In previous versions of Tristram, I declared a
>global array CHESS_MOVE moves[MAXPLY][MAXPLY].
>
>This time I'm thinking about using a doubly linked
>list, or even an array of singly linked list
>MoveList moves[MAXPLY].
>
>This means that every time I wanted to add a move to
>one of the list, I'd have to call the "new" routine
>to allocate the memory.
>
>Before I get into it, has anybody tried this before
>and found it too expensive?
>
>--
>James

You want to stay away from new anywhere in your search loop.  I started
with C++ and had never written code that needed to be fast before.  My
first implimentation allocated memory for each move.  It is especially
slow in the debugger.  I found another way.

Brian



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.