Author: Ernst A. Heinz
Date: 22:25:31 07/30/98
Go up one level in this thread
On July 30, 1998 at 14:58:56, Bruce Moreland wrote: > >On July 30, 1998 at 11:54:35, Frank Schneider wrote: > >>On July 30, 1998 at 10:03:27, blass uri wrote: >> >>>7k/7p/8/5p2/2K2P1P/8/8/8 white to move and win. >>> >>>The winning move Kc5. >>> >>>Fritz5 needs only 11 seconds to solve it on pentium200MMX >>>the same is for Fritz4. >>> >>>Junior5 needs more than 20 minutes to solve it in depth 54. >>> >>>how much time do other programs need? > >>Gromit X (980729) on a K6/200, 12MB hashtables >>needs 2:10 and finds Kc5 at ply 29. > >I would be curious to know how long they take to figure out that Kd5 draws. >Also, what the score jump when the thing switches from Kd5 to Kc5, which I >assume it will do. > >bruce "DarkThought" on a 600MHz Alpha-21164a with 4M transposition-table entries per side finds Kc5! in 8 sec. It switches to Kc5 in iteration #26 with a score of +3 and then fails high on it later (score >= 10). Up to iteration #26 it prefers Kd5 at roughly +2. =Ernst=
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.