Author: leonid
Date: 16:31:41 08/12/01
Go up one level in this thread
On August 12, 2001 at 18:20:20, Heiner Marxen wrote:
>On August 12, 2001 at 06:21:03, leonid wrote:
>
>>[D]rqn1QB2/r2nQ1K1/b2qQ2R/b3Q2B/q3N2R/q2qQ2N/q2qQ1Q1/qkq1QQ2 w - -
>>
>>Please indicate your result.
>>
>>Thanks,
>>Leonid.
>
>Hello Leonid,
>
>I'm afraid, your CD is a bit too hard for Chest: after 6.4 hours on a K7/600
>with 350MB hash it still just says "No solution in 9 moves."
>More depth will last a bit too long, so I stop here. Sorry.
Good! You went one move deeper that me. I stopped at move 8 where mine already
reached 3 hours and 5 min. Now only move ten is not certain for me. Mate in 11
was found by selective for sure.
Celeron 600. No hash.
>EBF is not very bad, but also not good enough:
>
> depth time EBF[T] EBF[N]
># 1 0.00s 0kN 0.87 1- 0
># 2 0.00s 0kN 1.00 1- 0
># 3 0.04s 2kN [ 19.16] 0.97 96- 0
># 4 0.53s [ 13.25] 24kN [ 12.18] 1.03 1698- 0
># 5 4.87s [ 9.19] 229kN [ 9.42] 1.16 18675- 0
># 6 42.69s [ 8.77] 1955kN [ 8.55] 1.30 188341- 0
># 7 313.41s [ 7.34] 14541kN [ 7.44] 1.56 1439982- 3
># 8 2708.83s [ 8.64] 127471kN [ 8.77] 1.76 12218709- 3595572
># 9 23079.04s [ 8.52] 1096507kN [ 8.60] 1.86 104265576- 95517675
Your initial speed say me once again that I should go and look very closely my
"special plys" writing. I do remember that few ideas about few plys above
actual specialized plys I rejected in the past as too clumzy. But, probably, I
should come back and insist. For now I am too far from programming.
My brute force data:
Moves Time Branching factor NPS
3 0.11 sec
15
4 1.648 sec 58k
8.7
5 14,34 sec 56k
8.73
6 125 sec 63k
9.8
7 20 min 26 sec 74k
9.04
8 3 h 4 min 40 sec 90k
As usual, deeper is the search, more is average NPS. Now I am almost sure that
selective search in mate solver is so good in providing very high NPS only
because search for mate stop mainly before ever reaching very specilized plys
layers. In not mate solving part of the program "specialized plys" are,
probably, not that efficent and they are used also more often that in mate
solver.
"CD" this time came to my mind just because this is one abreviation from
computer world.
Cheers,
Leonid.
>Cheers,
>Heiner
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.