Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Crafty null move question

Author: Robert Hyatt

Date: 09:16:55 07/18/02

Go up one level in this thread


On July 18, 2002 at 12:08:11, Steffen Jakob wrote:

>Hi Bob!
>
>On July 18, 2002 at 11:52:39, Robert Hyatt wrote:
>
>>>Calculating <421 is faster than calculating 7*INCPLY again and again.
>>
>>
>>Speed is not an issue for the calculation.  if INCPLY is a constant, then
>>the compiler realizes 7*INCPLY is _also_ a constant and will produce the code
>>using the constant 420 rather than doing the calculation.
>>
>>The reason it is a constant (magic number as Steffen said) is that as I tested
>>this, I played with a _lot_ of different values, and I found things like
>>6*INCPLY+30 to be tedius to type.  Once I settled on 7*INCPLY I should have
>>used that to make it more readable of course, and I will fix this (and others)
>>as I notice them.
>
>Because I had a look at SwapXray yesterday I remember that the direction numbers
>are "magic" too :-)
>
>Same for the usage of 'a&7', 'a>>3' or 'outside&192'.
>
>Another thing I remember is that the EXTENSIONS type from chess.h is unused.
>Same for FAIL_LOW_POS.


Extensions was a debugging tool.  IE I used to have a parallel array with
the PV in one element and the extensions for that ply in another.  Then when
I dumped the PV I could see which plies had extensions applied.  I took that
out a long time back to save the time of backing up the stuff when I really
didn't care except on rare occasions during debugging.

The rest are simply carry-overs.  To me a&7 is clearer than a&CONST where I
might not remember which bits are on in CONST...






>
>Greetings,
>Steffen.



This page took 0.01 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.