Strength on a 8 MHz 286

[Christopher Dorr]

Hi! I just had a quick question. I remember a list in CCR suggesting what
the rating differencex were between different computers with various
processors. Unfortunately, I can't find my copies.

My laptop computer is a 286 with 640KB RAM, running at 8 MHz. (Yes,
computers like this still exist! :) I was wondersing what the suggested
rating difference between this and a 486/66 is for various chess programs
(Specifically, Zarkov 2.61 and Fritz II).

Thanks,

Chris

[ChessMan]

Here are IM Larry Kaufman adjustments for different processors - yours is not listed
but I would double the adjustemt for the 16 mhz 286 or -360 points.

http://knoll.kih.no:8001/stud/mads2/chess/kaufman.htm

Mike

Here's the text info from the above page:

-- Last update 19/10-95 --

Adjustment Table for PC Processors

Mr. Larry Kaufman (Chess Computer expert) has estimated the rating adjustment for a given program when moving from one PC-processor to another (Intel- or Intel-compatible processors).

Note that the numbers are general, and the table is not valid for all programs. Some programs gain more points than others when the processing power is doubled, while some programs gain less. It's widely accepted that by doubling the speed of the processor you normally gain 70-80 rating points (the gain seems to decrease at higher strength-levels).

The table below uses the 486/66 as a reference. Hash-table size is 4 MB for all programs which use more than 640K (conventional memory).

The table should be interpreted this way:
Let's say you're running PawnPusher 7.0 rated at 1.600 ELO on a 8088 4.77 MHz. You're planning to purchase a Pentium 90 to boost the playing-strength of your program. What performance could you expect from PawnPusher running on your new hardware? Simple mathematics gives: (-(-385)+65) + 1.600 = 2.050 points! Wow! PawnPusher should theoretically play 450 points better than before!

Larry Kaufman's Adjustment TableCPU MHz Adjustment Pentium 90 +65 Pentium 66 +50 Pentium 60 +40 486/DX4 100 +20 486/DX2 66 0 486/DX2 50 -25 486DX/SX 33 -35 486DX/SX 25 -60 486SX 20 -80 386DX 40 -80 386DX 33 -95 386DX 25 -120386SX 25 -140 386SX 20 -160 386SX 16 -180 286 16 -180 286 12 -205 8088 10 -320 8088 4.77-385

<Picture>Go Back

Please mail your comments, complaints or suggestions to:
ma...@tott.kih.no
Copyright © 1995 Mads Brevik

---------------------------------------------------------------------

Note - the <picture> has the ratings -- sorry.

[Ed Parry]

In a previous article, crd...@iac.net (Christopher Dorr) says:

>mail
>From: crd...@iac.net (Christopher Dorr)
>Newsgroups: rec.games.chess.computer
>Subject: Strength on a 8 MHz 286
>Date: 28 Dec 1995 02:35:27 -0500
>Organization: Internet Access Cincinnati 513-887-8877
>Lines: 14
>Message-ID: <4bthbv$l...@little-miami.iac.net>
>NNTP-Posting-Host: little-miami.iac.net
>X-Newsreader: TIN [version 1.2 PL2]

>
>Hi! I just had a quick question. I remember a list in CCR suggesting what
>the rating differencex were between different computers with various
>processors. Unfortunately, I can't find my copies.
>
>My laptop computer is a 286 with 640KB RAM, running at 8 MHz. (Yes,
>computers like this still exist! :) I was wondersing what the suggested
>rating difference between this and a 486/66 is for various chess programs
>(Specifically, Zarkov 2.61 and Fritz II).
>
>
>Thanks,
>
>Chris

Hi Chris!

1) Okay, I yanked mine out - page 39 in vol 5 no 2 - There test base
computer is a 486 DX 66 - so the adjustment for the 286/8 would be -
subtract 340 points.

Zarkov 2.5 is rated at 2308 and Fritz 2 is rated at 2356 on the 486 dx
66. A little math would show this puts em in the 1950-2050 range. Still
not too bad for most of us average players. 8-)

MCHESS v1.71 used to kick my butt EASY on the old 286/8 (True IBM AT no
less) I used to run.

Ep