Author: leonid
Date: 13:50:39 01/22/01
Go up one level in this thread
On January 22, 2001 at 14:02:25, Gian-Carlo Pascutto wrote: >On January 22, 2001 at 12:15:02, leonid wrote: > >>Program is Sjeng 8.5? > >Yes > >>Is this Cyrix CPU 120 Mhz? > >Yes > >>500000 nodes is the total number of nodes seen before finding the result? > >No. That is the maximum size the search tree can be. Sjeng's matefinder >stores _all_ nodes into memory. You can make the tree as large as your >available memory. 500000 nodes equals about 16M of RAM, which was the >limit I had set for this try. > >The actual number of nodes searched was 404050. > >>>P: 0 D: 100000000 N: 404050 S: 404049 Mem: 10.79M Iters: 4151 >>>Time : 172.480000 >> >>172 seconds? > >Yes. > >> >>>This position is WON. >> >>In how many moves? > >No way to tell for sure. The algorithm is highly selective best-first >so there is no guarantee whatsoever that what it finds is the shortest >mate. You could establish a lowerbound by checking the maximum depth it >reached, but it follows forced lines very deeply... Thanks for responding! It was very interesting to read. I am afraid that my last question was not correctly understood. Maybe my way of saking was not clear. It is possible. I wanted to know how many moves deep program looked before finding this mate. I had the impression that your program just goes automatically searching for the mate, never mind how deep it is. I am wrong? Leonid. >GCP
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.