On Wed, 02 Aug 2000 22:41:24 -0700, Rick wrote:
> So you think those "anomalies" are related to either processor speed or
> RAM or ... I'll have to load "Rachne" up on this (450AMD/128Mb) or the
> (486DX66/32Mb) to check that out. Or might that be the old video card?
>
Hope you don't mind me chipping in here! I've wondered about strange
differences between seemingly identical machines for a while now. Whilst
my experiment didn't directly involve Arachne, I hope that it might be
relevant:
I wrote a little programme in Turbo Basic to do FFT nad then draw a
few lines on the screen to represent the spectrum plot. The graphics
were hardly stunning. TB only just recognises VGA! It also read the
timer before and after so that it would print on the screen how long the
process took. My main machine is a 486DX4-100 VLB (with Diamond VLB
video) and I was getting times of around 3 seconds. I installed the
programme on my other machine. This is also a 486DX4-100, same HD type
and RAM size (16Mb) and it took just over 0.5 seconds!!! The ONLY
difference I could see was that the other 486 is a PCI bus with PCI
video. Both machines had DOS 5.00, too - also CPU cache turned on, etc.,
etc. HD isn't used once the programme is loaded, so it can't be that -
anyway, the HDs are similar.
Must be the video, do we think?
The really interesting bit from this list's point of view is that the
'other' 486 just wouldn't run Arachne a few versions back. It knew where
the COM port was and would browse files off-line a treat, but simply
refused to connect on-line after dialling. In desperation, I even tried
swapping the CPUs between machines but it didn't make any difference.
I must admit that I haven't yet got around to trying V1.66 on it, but
maybe I should, in the light of these 'difference' discussions....
Finally, I for one don't think that I've found any significant
speed improvement in V1.66 over V1.64, using my main machine.
Isn't this testing such FUN!
Ron.
-- Arachne V1.66, NON-COMMERCIAL copy, http://arachne.cz/
