Author: Peter Kappler
Date: 19:51:29 05/30/99
Go up one level in this thread
On May 30, 1999 at 22:17:05, Daniel Karlsson wrote: >The following is a statistics dump from my hashtable after searching a test >position (Bratko-Kopek 4 - and no, I didn't find the best move :-( ). > >CHashTable dump: >131072 entries in 3072KB (53.5% used) > 610 (0.9%) exact > 31577 (45.1%) upper bounds > 37890 (54.1%) lower bounds > 0 (0.0%) stale >1448685 probes > 140246 hits > 1308439 misses >100229 writes > 23560 collisions > >Does any of this look strange? I'm a bit worried about the low hit ratio and >high collision frequency but maybe that's normal? Any input would be >appreciated. > >BTW I'm using a single 'depth preferred' hash table. I count a probe whenever I >ask the hashtable if it has a score for the current position and depth, and a >hit when it answers 'yes'. A 10% hit ratio sounds about right for a middlegame position. I assume by "hit" you mean that you were able to use the score that was returned. Could you post the EPD? Actually, could you post the whole suite -- mine seems to be corrupted... --Peter
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.