Author: Vincent Diepeveen
Date: 03:43:53 11/13/99
Go up one level in this thread
On November 13, 1999 at 06:21:38, Frank Quisinsky wrote:
>On November 13, 1999 at 06:20:23, Frank Quisinsky wrote:
>
>>Hello,
>>
>>great idea !!!
>>
>>I hope SOS, Patzer, Comet, Phalanx , AnMon and other play this tournament !
>>I think a good time is 40 moves in 40 minutes !
60 seconds a move sucks bigtime Frank.
Don't need a tournament for that. Is already getting done at icc.
90 0 is already *very* quick. Against my unattended program some
attended will for sure forfeit and lose the game by means of that.
Especially DOS programs are losing quite some time.
Average losing 8 seconds a move.
Much better level is of course 40 in 2, but icc doesn't allow that.
ICC doesn't allow x moves in y time, which is indeed quite laughable.
What icc DOES allow is a level of for example 90 10.
This does give 90 minutes whole game and 10 seconds increment.
Even slow operators will not forfeit then when it's getting a
lot of moves.
I suggest modifying 90 0 to 90 10.
>>Kind regards
>>Frank
>This tournament without ponder !
>No, no a joke !
Yeah and then at a k6-450 or something...
no ponder k6-450 which is as slow as a PII400 or celeron 366 for diep...
Then also no ponder.
So at a dual 550 which is 1100Mhz together:
1100 / 366 = 3.0 times faster. I estimate permanent brain
at 40% speed win, so 3.0 x 1.4 = 4.2x faster.
Then secondly you run 2 programs on that 450K6
which means again that cpu time is divided ==> 8.4 times faster,
as even though a program is not searching PB moves, it's not idling, so
still eating 50% system time (in the case of DIEP at least).
So your 40 in 40 level testing is for the internet equal to game in 5
minutes.
I didn't take into account yet:
- lag
- hashtables, as 300mb hashtables, which is 10 times more than
at your K6, is quite some difference...
>Frank
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.