Testing Transposition Tables

Discussion of chess software programming and technical issues.

Moderator: Ras

User avatar
hgm
Posts: 28395
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: Testing Transposition Tables

Post by hgm »

Fairy-Max, which has not the most advanced of all hash tables (pure replace-always) takes 5 sec (on 1.3GHz Pentium-M) to find the promotion at 23 ply with 96MB hash table:

Code: Select all

28	+8.32	28.2M	0:43.41	e1f2
 27	+8.21	16.9M	0:26.11	e1f2
 26	+8.18	8.9M	0:13.79	e1f2
 25	+8.11	5.2M	0:07.98	e1f2
 24	+8.13	4.3M	0:06.68	e1f2
 23	+8.13	3.4M	0:05.24	e1f2
 22	+1.98	2.6M	0:04.07	e1f2
 21	+1.99	1.9M	0:02.97	e1f2
 20	+2.01	1.4M	0:02.19	e1f2
 19	+2.05	951045	0:01.47	e1f2
 18	+2.07	587444	0:00.90	e1f2
 17	+1.44	391121	0:00.61	e1f2
 16	+1.46	325660	0:00.51	e1f2
 15	+1.47	261747	0:00.41	e1f2
 14	+1.24	182258	0:00.29	e1f2
 13	+0.90	143694	0:00.24	e1f2
 12	+0.79	93208	0:00.15	e1d1
 11	+0.94	65344	0:00.11	e1d2
 10	+0.81	38008	0:00.07	e1f1
  9	+0.93	22228	0:00.05	e1f2
  8	+0.86	12762	0:00.02	e1d2
  7	+0.85	7621	0:00.01	e1f2
  6	+0.85	2476	0:00.00	e1f1
  5	+0.82	963	0:00.00	e1d2
  4	+0.86	591	0:00.00	e1d2
  3	+0.85	231	0:00.00	e2e4
  2	+0.74	54	0:00.00	e1d1
  1	+0.88	11	0:00.00	e1f2
  0	+0.74	2	0:00.00	e2e4