Chessmaster 4000T Performance Benchmark
Harold Jack Meadows
selection under the HELP menu that reads:
Performance Benchmark running on 486-66 DX/2
Personality: Chessmaster
Layout: Plain and Simple
Fixed Depth: 8 ply search
1. h2-h4 ... black move at 00:58
With my 486-66 SX, it took 01:39 to make the move! Quite dissapointing.
Does the math coprocessor in the DX/2 speed things up that much or is
there something else slowing things down? I have 8 Meg of RAM memory.
How fast is your computer with this benchmark? Should be interesting to
compare the relative speed of computer systems using this common benchmark.
--
_______________________________________________________________________________
Harold Jack Meadows jack...@netcom.com
NETCOM, the West Coast's Leading Internet Service Provider. (408) 554-8649
Michael Atkins
Dave Linder (d...@julian.uwo.ca) wrote:
: In article <jackmeadC...@netcom.com> (Harold Jack Meadows) writes:
: >In the Chessmaster 4000 Turbo for windows program, there is a "Chess Engine"
: >selection under the HELP menu that reads:
: > Performance Benchmark running on 486-66 DX/2
: > Personality: Chessmaster
: > Layout: Plain and Simple
: > Fixed Depth: 8 ply search
: > 1. h2-h4 ... black move at 00:58
: >With my 486-66 SX, it took 01:39 to make the move! Quite dissapointing.
: >Does the math coprocessor in the DX/2 speed things up that much or is
: >there something else slowing things down? I have 8 Meg of RAM memory.
: I got it down to 01:29 on my 486-40 DX with 8Mb RAM.
: If I have my barclock running, and flock wallpaper, it takes 01:31.
: If I also have Winsock running a SLIP connection it takes 01:33.
: It played 1...Nf6, for those who are dying to know...
: Dave
: d...@julian.uwo.ca
: ps. ...perhaps if I closed the Program Manager...?
: >How fast is your computer with this benchmark? Should be interesting to
Hugh Bothwell
>With my 486-66 SX, it took 01:39 to make the move! Quite dissapointing.
For Windows 3.1, you can specify how much time to give each running program (
size of time slice); if you aren't in a habit of writing memos while playing
chess, you can give Chessmaster more or ALL the time. I believe the setting
can be reached by selecting the Chessmaster icon, Edit Item (under File? I
think), Advanced Options. The time is specified in ms, and is usually 20;
setting it to something like 100 will slow other simultaneously
running programs down but should speed ChessMaster up considerably.
I'm not sure if it would work, but you could try running it in Standard Mode
as well - can only run one program at a time, but would really speed it up.
Start Windows by type "win -s".
=====================================================================
Hugh Bothwell ! ! F-14D Tomcat
bhbo...@HELIX.watstar.uwaterloo.ca | _ | by Aapool Biman
(519)885-5444 __|__(o)__|__
El. Eng Undergrad --------======/+-+/ . \+-+\======--------
University of Waterloo _____( |_|\___/|_| )_____
Ontario, Canada \__/ * * \__/
/ \
Rap Music: bad poetry to a continuous drum solo.
Dave Linder
>selection under the HELP menu that reads:
> Performance Benchmark running on 486-66 DX/2
> Personality: Chessmaster
> Layout: Plain and Simple
> Fixed Depth: 8 ply search
> 1. h2-h4 ... black move at 00:58
>With my 486-66 SX, it took 01:39 to make the move! Quite dissapointing.
>Does the math coprocessor in the DX/2 speed things up that much or is
>there something else slowing things down? I have 8 Meg of RAM memory.
I got it down to 01:29 on my 486-40 DX with 8Mb RAM.
If I have my barclock running, and flock wallpaper, it takes 01:31.
If I also have Winsock running a SLIP connection it takes 01:33.
It played 1...Nf6, for those who are dying to know...
Dave
d...@julian.uwo.ca
ps. ...perhaps if I closed the Program Manager...?
Harold Jack Meadows
: In article <jackmeadC...@netcom.com>
: jack...@netcom.com "Harold Jack Meadows" writes:
: >The Chessmaster 4000 Turbo for windows program, has a "Chess Engine"
: > selection under the HELP menu that reads:
: > Performance Benchmark running on 486-66 DX/2
: > Personality: Chessmaster
: > Layout: Plain and Simple
: > Fixed Depth: 8 ply search
: > 1. h2-h4 ... black move at 00:58
: >
: > With my 486-66 SX, it took 01:39 to make the move!
: >
: I would have thought
: that a SX2/66 would have been very similar in time to a DX2/66 as I
: would think that CM4000 wouldn't use a maths co-pro. (how many woulds
: can I get in one sentence!) When running the benchmark nothing else
: must be running e.g. third party shells etc.
: Hope this helps
: Al
I called COMPAC technical service and the representative says: "The reason
your SX/2 is 1.7 times slower than a DX/2 is because the SX doesn't have
a math coprocessor....yes, even though CM4000 Turbo does not use any
floating point math in it's algorithm!" Hummmmm the coprocessor must
speed up the graphics or something like that.
There is a Pentium 90 somewhere in my future.
Harold Jack Meadows
: >With my 486-66 SX, it took 01:39 to make the move! Quite dissapointing.
: For Windows 3.1, you can specify how much time to give each running program (
: size of time slice); if you aren't in a habit of writing memos while playing
: chess, you can give Chessmaster more or ALL the time. I believe the setting
: can be reached by selecting the Chessmaster icon, Edit Item (under File? I
: think), Advanced Options. The time is specified in ms, and is usually 20;
: setting it to something like 100 will slow other simultaneously
: running programs down but should speed ChessMaster up considerably.
I tried this....but no improvement.
: I'm not sure if it would work, but you could try running it in Standard Mode
: as well - can only run one program at a time, but would really speed it up.
: Start Windows by type "win -s".
CM4000 Turbos would not run in Standard Mode.
Thanks for the good suggestions, Hugh, but unfortunately the bottleneck is
somewhere else. According to COMPACT, a math coprocessor is needed for
best performance even though CM4000 doesn't use any floating point math
in the algo...hummmm.
Randell Jesup
>: >
>: > With my 486-66 SX, it took 01:39 to make the move!
>I called COMPAC technical service and the representative says: "The reason
>your SX/2 is 1.7 times slower than a DX/2 is because the SX doesn't have
>a math coprocessor....yes, even though CM4000 Turbo does not use any
>floating point math in it's algorithm!" Hummmmm the coprocessor must
>speed up the graphics or something like that.
Hogwash.
The probable reason is a worse memory system, especially less cache.
Second most likely factor is Windows configuration/TSR's/background programs.
Remember, a computer is more than a CPU!
--
Randell Jesup, Scala US R&D
Randel...@scala.com
Ex-Commodore-Amiga Engineer, class of '94
#include <std/disclaimer>
Ariel Cohen
>: > Performance Benchmark running on 486-66 DX/2
>: > 1. h2-h4 ... black move at 00:58
>: >
>: > With my 486-66 SX, it took 01:39 to make the move!
>I called COMPAC technical service and the representative says: "The reason
>your SX/2 is 1.7 times slower than a DX/2 is because the SX doesn't have
>a math coprocessor....yes, even though CM4000 Turbo does not use any
>floating point math in it's algorithm!" Hummmmm the coprocessor must
>speed up the graphics or something like that.
Bull. Making the move takes 01:37 on my 486SX-33; if it takes 01:39 on your
SX2-66 then you have a very lousy setup...
BTW, has anyone tried this on a DX4-100 or an AMD DX2-80?
Ariel
TIMOTHY OLSON
> Ariel
the 66 MHz chips sound like clock doubled 33's. if that's that case,
then the bus on is likely to still run at 33 MHz, so anytime info comes
off the chip, it has to slow down to 33. it may be that the DX/2 has a wider
channel motherboard or that the board's bus simply runs faster. i
believe that this difference in bus width/speed was the difference
between 386 DX's and SX's. this still may be true in addition to the
486 SX's lack of a math cope.
HIARCS 3
writes:
I have tried it on a Dell 90 mhz pentium - factory configured - 28
seconds was the time.