Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Does this look "right" to you?

Author: Robert Hyatt

Date: 08:36:46 08/12/03

Go up one level in this thread


On August 12, 2003 at 02:55:11, Rick Bischoff wrote:

>5 -18 0 6812 g1h3 d7d5 h3g5 e7e5 d2d4
>5 -16 0 11159 b2b3 d7d5 d2d4 b8c6 b1c3
>5 3 0 16025 c2c3 d7d5 d2d4 b8c6 g1f3
>5 17 0 21141 d2d3 d7d5 e2e4 e7e5 g1f3
>5 30 0 25914 e2e3 d7d5 d2d4 b8c6 g1f3
>5 35 0 33920 d2d4 d7d5 b1c3 b8c6 g1f3
> ab returned: 35
> probes: 26758 hits: 11144 return bad: 23202 hashhitbutflagbad: 0
>hashhitbutdepthbad: 5 overwritesamekey: 5 hashoverwriteold: 19 hashstore: 23202
>15595 unique keys in table
>
>
>This is the output from running a fixed-depth search at the starting
>position....
>probes = Number of times the probe function was called
>hits = Number of times the entry in the hashtable had the same key as the
>current position
>bad = Number of times "probe" could not  return a good value
>hashhitbutflagbad = Number of times we had a hit but had the flag set to bad
>hashhitbutdepthbad = Number of times we had a hit but insuffiecent depth
>
>hashstore = Number of times store was called
>overwritesamekey = Number of times we overwrote the same key (if depth >
>stored_depth)
>hashoverwriteold = Number of times we overwrote a different key (ignoring empty
>entries)
>
>The probes/hits seems a little high to me, but I haven't enough experience to
>know for sure.  Anyone care to speak up?
>
>Rick


Looks high to me as well.  However, the "bad" number drops the effective
hits to 10%, which seems reasonable.  You might check your "hit" counter
to see if it is really being updated right.



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.