Author: Christophe Theron
Date: 13:19:51 03/26/98
Go up one level in this thread
On March 26, 1998 at 08:36:39, Kai Lübke wrote:
>Could you post/mail the full LCT results for Chess Tiger?
>I will include it on my website.
Here they are:
Computer is K5-100MHz, 256Kb L2 cache, 16Mb hash tables. This CPU is
roughly equivalent to P200MHz, and EXACTLY half as fast as a K6-200MHz
when running Chess Tiger. So I will also give an approximated result for
K6-200.
Test were done in a W95 DOS Box (would be about 3% faster if run under
real DOS).
Chess Tiger 11.4 gets an elo estimation of 2530 on K5-100, 2590 on
K6-200. Here are the details, position by position. Time is given for
K5-100.
Position Time (s) Points
------------------------
POS01 000.49 30
POS02 000.06 30
POS03 285.06 10 +5 on K6-200
POS04 019.55 25 +5 on K6-200
POS05 112.54 15 +5 on K6-200
POS06 054.81 20 +5 on K6-200
POS07 588.25 05 +5 on K6-200
POS08 005.16 30
POS09 010.16 25 +5 on K6-200
POS10 Not found in 10mn
POS11 Not found in 10mn
POS12 Not found in 10mn
POS13 Not found in 10mn +5 on K6-200 (632s on K5-100)
POS14 Not found in 10mn
TOTAL POS = 190 / 420 (45%)
CMB01 000.65 30
CMB02 000.33 30
CMB03 002.47 30
CMB04 008.57 30
CMB05 002.36 30
CMB06 024.12 25
CMB07 077.06 20
CMB08 010.32 25 +5 on K6-200
CMB09 009.44 30
CMB10 137.25 15 +5 on K6-200
CMB11 Not found in 10mn
CMB12 Not found in 10mn
TOTAL CMB = 265 / 360 (74%)
FIN01 000.32 30
FIN02 000.06 30
FIN03 020.27 25
FIN04 044.38 20 +5 on K6-200
FIN05 008.51 30
FIN06 012.08 25 +5 on K6-200
FIN07 179.61 15 +5 on K6-200
FIN08 Not found in 10mn
FIN09 Not found in 10mn
TOTAL FIN = 175 / 270 (65%)
ELO on K5-100MHz = 1900 + 190 + 265 + 175 = 2530
There are at least 12 positions that would get 5 more points on K6-200.
ELO on K6-200 = 2590
Here is how the points are given in LCT:
0 - 9.99 s -> 30 pts
10 - 29.99 s -> 25 pts
30 - 89.99 s -> 20 pts
90 - 179.99 s -> 15 pts
180 - 389.99 s -> 10 pts
390 - 599.99 s -> 5 pts
10mn and above -> 0 pts
The "Elo base" is 1900. Times recorded are the last time the program
changed and choosed the right key move, without changing its mind until
10mn.
Yesterday I wrote that CT got 270 points in the CMB part. I had booted
directly in DOS, and CMB08 was under 10s (maybe 9.95 or so). A 5 points
difference doesn't matter, anyway.
I am not claiming that CT11.4 is really worth 2590 points on K6-200. My
guess is that against other programs it would deserve between 2400 and
2500 on this CPU. And it still fails on the advanced positional tests.
Christophe
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.