Computer Chess Club Archives


Search

Terms

Messages

Subject: Thanks (nt)

Author: Mike Siler

Date: 08:31:56 12/13/03

Go up one level in this thread


On December 13, 2003 at 11:30:37, Albert Bertilsson wrote:

>On December 13, 2003 at 11:28:04, Mike Siler wrote:
>
>>My perft function is undercounting by a few nodes every time and I can't figure
>>out what it is not counting. When I do perft 5 it finds 4863894 nodes as opposed
>>to the correct number 4865609. It also finds the following stats
>>
>>0 castles
>>0 promotions
>>83981 captures
>>258 en passants
>>
>>For each first move it gets the following number of nodes:
>>
>>a2a3 181037
>>a2a4 217823
>>b2b3 215246
>>b2b4 216098
>>c2c3 222852
>>c2c4 239989
>>d2d3 328461
>>d2d4 361492
>>e2e3 402525
>>e2e4 404884
>>f2f3 178842
>>f2f4 198428
>>g2g3 217201
>>g2g4 214001
>>h2h3 181035
>>h2h4 218820
>>b1a3 198563
>>b1c3 234620
>>g1f3 233484
>>g1h3 198493
>>
>>I can't think of what else to list here. I'd appreciate it if someone would do
>>me the favor of checking these against their numbers to help me find what I'm
>>not counting.
>>
>>Michael
>
>Sharper has a divide command producing a list simmilar to yours, use that to
>compare for which move your movegen is wrong. Do that move and repeat the
>process, that should let you narrow down the problem very fast.
>
>/Regards Albert



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.