Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!rutgers!att!linac!uwm.edu!ux1.cso.uiuc.edu!uxa.cso.uiuc.edu!gcw20877
From: gcw20877@uxa.cso.uiuc.edu (George Wang)
Newsgroups: comp.sys.laptops
Subject: Re: Dataworld LP-320 analysis
Message-ID: <1990Dec9.000752.9005@ux1.cso.uiuc.edu>
Date: 9 Dec 90 00:07:52 GMT
References: <1990Dec4.223137.23174@aeras.uucp> <1990Dec6.164923.5100@ux1.cso.uiuc.edu> <1990Dec7.162834.24566@csn.org>
Sender: news@ux1.cso.uiuc.edu (News)
Organization: University of Illinois at Urbana
Lines: 28

In article <1990Dec7.162834.24566@csn.org> atk@tigger.Colorado.EDU (Alan T. Krantz) writes:
>>>The anomalies are why a 16 Mhz 386 has a higher compute index
>>>than a 20 Mhz 386, and why there was an increase in performance
>>>in one of the CI results on the LP-320.
>>I too have noticed this!! I suspect that the internal bus may
>>be limited somehow in the laptop.. I think the motherboard may
>>have been designed non-optimally.. I've got a 5.25" 1.2 meg drive
>
>What is probably more likely is that this 20 Mhz system has 1 or 2 
>wait-states and the 16 Mhz has 0 or 1 wait states.  This info should
>be in the documentation of both computers. If not you can probably 
>figure it out by the speed of the memory (just look at the chips)...
>
>I think a 20mhz would have to have 1 wait state without a cache (using
>common 1mb chips) and the 16mhz could just squeeze by with 0 wait
>states using common 1mb chips (doubled banked) - but I forget these things...

The Zeos 386DX-20 is a ZERO wait state machine... The pheonix
bios lets you change the wait states but the si indexes I quoted
were for a 20 mhz ZERO wait state with a RAS Precharge of 3 CLK2...
Anyone else got other ideas on why the zeos motherboard is slower than regular desktop
boards?
--
George Wang
University of Illinois at Urbana Champaign
(217) 332-4019
INTERNET: gcw20877@uxa.cso.uiuc.edu
UUCP: gargoyle!igloo!gwang