Computer Chess Club Archives


Search

Terms

Messages

Subject: hash collisions

Author: martin fierz

Date: 05:13:52 01/31/01


hi,

i recently corrected some code in my connect 4 program and now it is able to
solve connect 4 in less than a day on a fast PC with a large hashtable (of
course, connect 4 has been solved long ago). i tried again with a smaller
hashtable and
got some strange results, for very long searches (billions of nodes) i don't get
the same value for
the root position. for not-so-deep searches i get the same values for both
versions. i am wondering, [if this is not just a bug :-)] could this be some
hashcollision-problem? can anybody give me a probability for a hash collision
occuring when using B-byte key & lock, with a hashtable with S entries after
searching N nodes? (i am using 4 byte ints for the key and the lock)
also, i think i remember somebody mentioning here that one can choose the random
numbers for the XORs in a clever way making hashcollision probabilities
smaller - can somebody tell me how?

cheers
  martin





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.