Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Crafty and NUMA

Author: Vincent Diepeveen

Date: 06:08:46 09/03/03

Go up one level in this thread


On September 03, 2003 at 08:12:55, Uri Blass wrote:

>On September 03, 2003 at 02:24:00, Gian-Carlo Pascutto wrote:
>
>>On September 02, 2003 at 22:34:49, Robert Hyatt wrote:
>>
>>>>Been working a year fulltime now :)
>>>>
>>>
>>>So?  It took you over a year to get your parallel search working.  It took
>>>me weeks.
>>>
>>>:)
>>
>>In all fairness, he did a full DTS implementation, including rewriting the
>>program to a nonrecursive search, while you took an easy way out.
>
>I do not understand the need for non recursive search.

So why don't you skip joining this discussion. Until you implement a parallel
search yourself you won't see the use either. It took some commercial
programmers weeks to realize why. Some of them already parallel.

>I think that non recursive search simply limit your possibilities for future
>developement because the code is ugly and you need to write almost the same
>function again and again.

Do you really believe that what the compiler can do can't be efficiently done in
your own software either?

>If you want to change something in the search rules then you need to change your
>program in a lot of places.

Why would that be, because you are not such a good programmer i guess?

Why would it be any different?

Let's talk about the good aspects, i save out a lot of expensive function calls.

If i remember well The King isn't recursive either. Johan no doubt reads this
and will say: "no way" when it isn't :)

I am sure fritz is non recursive too, because calling a function each time is
just too expensive for Frans.

>I guess that you need to write code for every possible depth that you get and in
>order to let your self to do extensions you need to write code for
>depth 10,depth 10 after one extension,depth 10 after 2 extensions, and you also
>need to limit the number of extensions at specific depth.
>
>You also limit your possibilities to extend because
>you cannot decide to extend more than one ply without modifying your code.

i can extend way easier of course, i can even extend looking back at previous
recursions. You cannot unless you have saved all that information in arrays.

I don't need to store it double then. you do.

><snipped>
>>
>>Diep's parallel performance does seem to be better than what you and I are
>>getting.
>
>I have no idea about Diep's parallel performance.
>I do not know about a single game of Diep on the new machine and I guess that we
>need to wait for november to see its performance.
>
>Uri

And i'm not going to post anything either till then.



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.