Author: Daniel Clausen
Date: 09:03:52 09/16/03
Go up one level in this thread
On September 16, 2003 at 11:51:15, Anthony Cozzie wrote:
>On September 16, 2003 at 11:33:53, Daniel Clausen wrote:
>
>>On September 16, 2003 at 10:56:49, Anthony Cozzie wrote:
>>
>>>I am in the process of restructuring Zappa's datatypes, moving from
>>>
>>>Search(varlist) {
>>> local_decls;
>>> .....
>>>
>>>to Search(RecursionBlock *rb) {
>>> ....
>>>
>>>where a RecursionBlock holds all the information that used to be local.
>>>
>>>I think there are 3 advantages to this:
>>>
>>>1. Converting to an iterative search would be much easier.
>>
>>True. And I assume that's your main goal?
>
>at some point i want to do a full DTS implementation.
I see.
>>>2. All data is accessible at any time/place in the search.
>>
>>Funny. I would have listed that under "disadvantages". :)
>
>you c++ abstraction wussies. IMO, making a chess engine fast is more important
>than abstraction. Its not like "int alpha" is going to change its function any
>time soon.
Hehe! I'd also put it under "disadvantages" in pure C. :) And no, "int alpha"
probably isn't going to change any time soon, but I would hate it, if the
evaluation function not only evaluates the current position but destroys the
movelist 2 plies higher in the tree. :))
>for example trans_table_lookup() requires, alpha, beta, ply, depth, etc.
>How do you get around this?
I would pass alpha, beta, ply and depth as parameters to trans_table_lookup().
:) Do you feel that 4 parameters are too much? (I guess you have speed
concerns?) IIRC, the search() of my C-engine had 8 parameters. That's probably
the limit I'd like to have. Otherwise I feel that the function is too
complicated anyhow.
>and if you tell me "class RecursionBlock" I *will* smack you ;)
Lol!
Sargon
PS. Sheesh, if I invested my time in the coding of my engine instead of being
puristic, I'd beat all you guys! ;) [except Diep, of course]
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.