Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!corton!chorus!opera.chorus.fr!mir
From: mir@opera.chorus.fr (Adam Mirowski)
Newsgroups: comp.sys.ibm.pc.hardware
Subject: Re: ET4000 speed
Message-ID: <8103@chorus.fr>
Date: 5 Mar 91 12:48:45 GMT
References: <1142@applix.com>
Sender: mir@chorus.fr
Reply-To: mir@opera.chorus.fr (Adam Mirowski)
Organization: Chorus systemes, Saint Quentin en Yvelines, France
Lines: 27

In article <1142@applix.com>, scotte@applix.com (Scott Evernden) writes:
%% Could someone with an ET4000 (e.g., Orchid ProDesigner II) and the
%% vidspeed benchmark (or something which measures transfer speed, like
%% Manifest), please send me a run of the benchmark?  
%% 
%% I am beginning to believe I should be getting better thruput than the
%% 2200k bps I'm seeing right now...

With the Landmark, I get 2.943 Mb/s on my Trident 8900. I run my
bus at 11 MHz.  On 8.33 MHz, I used to get 2.1-2.2 Mb/s. Also, when
I have put the card into 8 bits bus mode, the speed was about 1.45
Mb/s (I think the bus was already at 11 MHz).

I also tried to setup the jumpers to say I only have 256 Kb of RAM.
Normally, I have four 4256-80 chips which give me 512 Kb, and four
sockets are free. As I expected, the speed was 200-300 Kb/s lower.
The card must use a 16-bit internal data path when at least 4 RAM
chips are present (4 x 4 bits/chip) and multiplexes external 16
bits data accesses into 8 bits accesses when only 2 RAM chips are
there.

During that process, from time to time the 8900 must need additional
time, so it asks the bus for supplemental wait-states. That is why
the speed is slightly slower with an internal 8 bit data path.
-- 
Adam Mirowski,  mir@chorus.fr (FRANCE),  tel. +33 (1) 30-64-82-00 or 74
Chorus systemes, 6, av.Gustave Eiffel, 78182 Saint-Quentin-en-Yvelines CEDEX