Author: Ed Schröder
Date: 22:05:22 03/22/99
Go up one level in this thread
>Posted by Ernst Walet on March 22, 1999 at 16:12:00:
>>The hash table size in the CHESS 2010 experiment is indeed a very
>>unclear item.
>>
>>For the CHESS 2010 experiment I use 4 x Pentium-II 266, 64 Mb PC's
>>and 2 x Pentium-II 450, 128 Mb.
>>
>>My guess is that for some programs this is not enough for the 40/40:00
>>time control. Heck, I can't even tell for my own program.
>>
>>The "2010" formula is more or less based on the fact that every year
>>(single) processors become faster with a factor of about 0.75. Memory
>>will also go up nobody knows with how much but it will have a positive
>>effect to the speed increase too. Add up the software improvements
>>and I do think the factor 20 I have chosen is not such a bad one.
>>
>>Summery, just guessing...... and for what it is worth.......
>>
>>- 11 years x 0.75 gives a Pentium at 4000 Mhz (8 times faster)
>>- Hash table 64/128 Mb --> 1/2/4 giga byte --> 50% speed increase? (makes it
>>factor 12)
>>- More intelligent chess software --> factor=20
>>
>>Anyway these were my thoughts to play these games at the 40/40:00 level
>>to emulate tournament games (40/2:00) in the year 2010.
>>
>>There is another BIG unknown item. IMO we are not far away from the
>>point that multi-processor PC's will become standard and general available
>>at normal prices. If that is true chess programs enter a whole new area.
>>
>>Ed Schroder
>
>Dear Ed,
>
>I have two additional questions.
>
>1. Shouldn't you give the 266MHz/64MB machines a longer time control to
>make up for the speed difference with the 450MHz/128MB PC's?
One game already takes one week average to finish.
>2. It looks like you use auto232 for these games, does this mean that
>you have solved the problems with auto232?
No I don't have solved the problem. In fact I have dropped all DOS
auto232 testing to improve Rebel. I will try again with the Windows
version because the free auto232 Windows code looks much more
reliable than its older DOS brother.
I still don't have a clue what's going on with the DOS auto232 driver. I
have a long list of suspect things that proof something is very wrong.
The final knock came a few weeks ago when Rebel10c played an
auto232 match against Mchess8, level 40/40. To my astonishment
Rebel scored 19-2 which made me very suspicious as this simply
is unreal.
I was just going to abort this match because of the silly result when
my eye felt on the board position of the last game. Since Rebel10's
automatic annotation was on I noticed the two "??" on Mchess8 last
move given by Rebel.
When I took back the last move Rebel showed a draw score and
then the next move a +5.xx score. Have a look at the game yourself
and notice that 45.Ke1??? is a horrible blunder that loses immediately.
When I logged off of auto232 and started Mchess8 manually there is
NO sign of 45.Ke1 and when I force 45.Ke1 Mchess8 sees a +7.xx
score immediately.
The fact such a thing can happen made me drop all auto232 testing.
As for CHESS 2010 I have to use auto232. After each game I reboot
the machines and hope for the best.
Ed Schroder
Here is the raw auto232 output from Rebel.
[Event "?"]
[Site "1 min\move, P6 259 Mhz."]
[Date "1999.04.12"]
[Round "?"]
[White "Player"]
[Black "REBEL=BLACK"]
[Result "0-1"]
[WhiteElo "?"]
[BlackElo "?"]
[ECO "D58"]
1. d4 d5 2. c4 e6 3. Nc3 Nf6 4. Bg5 Be7 5. e3 O-O 6. Nf3 h6 7. Bh4 b6 8.
Be2 Bb7 9. Bxf6 Bxf6 10. cxd5 exd5 11. O-O c6 12. b4 a5 13. b5 c5 14. Rc1
Nd7 15. dxc5 Nxc5 16. Nd4 Qd6 17. Bg4 g6 { 00:01:29 9.00 0.07 g7-g6
Qd1-d2 Bf6-e5 Nd4-f3 Be5xc3 Qd2xc3 f7-f5 (c) } 18. Na4 Rfe8 { 00:01:04
9.00 0.22 Rf8-e8 h2-h3 h6-h5 Bg4-f3 Ra8-c8 Na4-c3 Nc5-e4 (c) } 19. Nxc5
bxc5 { 00:00:23 9.00 -0.07 b6xc5 Nd4-b3 Bf6-b2 Nb3xc5 Bb2xc1 Nc5xb7
Qd6-c7 (c) } 20. Nb3 c4 { 00:00:20 11.01 0.07 c5-c4 Rc1xc4 Bf6-e5 Rc4-c1
Be5xh2+ Kg1-h1 Bh2-e5 (c) } 21. Rxc4 Be5 { 00:00:00 10.08 0.10 Bf6-e5
Rc4-c1 Be5xh2+ Kg1-h1 Bh2-e5 a2-a4 f7-f5 (c) } 22. Ra4 Bxh2+ { 00:02:08
10.00 0.06 Be5xh2+ Kg1-h1 Bh2-e5 Ra4xa5 f7-f5 Bg4-h3 Ra8-c8 (c) } 23.
Kh1 Be5 { 00:01:44 10.00 -0.13 Bh2-e5 Ra4xa5 f7-f5 Bg4-h3 Ra8-c8 Ra5-a4
Rc8-c7 (c) } 24. Bf3 Qb6 { 00:02:09 9.02 0.00 Qd6-b6 Qd1-d3 Ra8-c8 Ra4xa5
Rc8-c3 Qd3-d2 Rc3-c4 (c) } 25. Qe2 Rac8 { 00:02:23 9.00 0.12 Ra8-c8
Rf1-b1 Rc8-c3 Ra4-h4 g6-g5 Rh4-a4 Re8-c8 (c) } 26. Kg1 Bc3 { 00:00:55
9.01 -0.05 Be5-c3 Rf1-d1 Re8-d8 Ra4-h4 g6-g5 Rh4-a4 f7-f5 (c) } 27. Rc1
Bb4 { 00:00:58 9.01 -0.09 Bc3-b4 Rc1-d1 Re8-d8 a2-a3 Bb4-c3 Ra4-h4 g6-g5
(c) } 28. Rd1 Re5 { 00:00:00 10.00 -0.12 Re8-e5 a2-a3 Bb4-c3 Rd1-c1 d5-d4
Bf3xb7 Qb6xb7 (c) } 29. a3 Bc3 { 00:00:00 10.00 -0.09 Bb4-c3 Ra4-h4 g6-g5
Rh4-h1 f7-f5 Qe2-d3 g5-g4 (c) } 30. Rc1 d4 { 00:01:04 9.00 0.03 d5-d4
Bf3xb7 Qb6xb7 Nb3xd4 Bc3xd4 Rc1xc8+ Qb7xc8 (c) } 31. Bxb7 Qxb7 { 00:00:00
10.00 -0.23 Qb6xb7 Nb3xd4 Bc3xd4 Rc1xc8+ Qb7xc8 Ra4xd4 Qc8-c1+ (c) }
32. Nxd4 Bxd4 { 00:00:14 10.00 -0.13 Bc3xd4 Rc1xc8+ Qb7xc8 Ra4xd4 Qc8-c1+
Rd4-d1 Qc1xa3 (c) } 33. Rxc8+ Qxc8 { 00:00:00 11.00 -0.13 Qb7xc8 Ra4xd4
Qc8-c1+ Rd4-d1 Qc1xa3 Rd1-b1 Re5-e4 (c) } 34. Rxd4 Qc1+ { 00:01:26 11.00
-0.03 Qc8-c1+ Rd4-d1 Qc1xa3 Rd1-b1 Qa3-e7 b5-b6 Qe7-b7 (c) } 35. Rd1
Qxa3 { 00:00:43 11.00 -0.04 Qc1xa3 Rd1-b1 Re5-e4 Qe2-d1 Qa3-e7 b5-b6
Re4-b4 (c) } 36. Rb1 Qe7 { 00:00:31 11.01 -0.11 Qa3-e7 Qe2-f3 Re5-e4
b5-b6 Qe7-b7 Rb1-b5 f7-f5 (c) } 37. b6 Qb7 { 00:00:51 11.00 -0.08 Qe7-b7
Qe2-c4 Kg8-h7 Rb1-c1 a5-a4 Qc4-a2 Re5-e7 (c) } 38. Qc4 Kh7 { 00:00:00 12.00
-0.10 Kg8-h7 Rb1-c1 a5-a4 Qc4-a2 Re5-e7 Qa2-a3 Qb7-e4 (c) } 39. Qa4 Rd5
{ 00:01:01 10.00 0.00 Re5-d5 f2-f3 f7-f5 e3-e4 f5xe4 Qa4xe4 Rd5-d7 (c)
} 40. f3 Re5 { 00:00:22 11.06 -0.17 Rd5-e5 e3-e4 Re5-c5 Qa4-a2 Rc5-h5
Qa2-a3 Rh5-e5 (c) } 41. e4 Rc5 { 00:00:24 11.00 -0.22 Re5-c5 Qa4-a3 Rc5-e5
Rb1-b2 Re5-h5 Qa3-d6 a5-a4 (c) } 42. Qa3 Re5 { 00:00:00 11.02 -0.34 Rc5-e5
Qa3-d6 Re5-g5 Rb1-b2 Rg5-h5 Qd6-d4 Qb7-b8 (c) } 43. Qc3 Rg5 { 00:01:14
11.00 -0.44 Re5-g5 Rb1-b3 Rg5-h5 g2-g4 Rh5-g5 e4-e5 a5-a4 (c) } 44. Kf2
Rh5 { 00:01:21 11.00 -0.29 Rg5-h5 Rb1-b2 Rh5-g5 Qc3-d4 Rg5-h5 Rb2-b3
Rh5-g5 (c) } 45. Ke1 Rh1# { 00:01:33 12.00 5.62 Rh5-h1+ Ke1-f2 Rh1xb1
Qc3xa5 Qb7-c6 Qa5-d5 Qc6xb6+ (c) } 0-1
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.