From: "Deranged Mutant" <WlkngOwl@unix.asb.com>
Date: Sat, 23 Mar 1996 11:07:58 +0800
To: coderpunks@toad.com
Subject: Maybe not (was "PC: Using BIOS Wait function as a source of entr
Message-ID: <199603220446.XAA21329@unix.asb.com>
MIME-Version: 1.0
Content-Type: text/plain



I did some more tests with the variation when sampling the timer 
before and after a call to the BIOS Wait function on my PC.  Set the 
delay to 977ms, which is allegedly the counter's resolution and 
collected samples at even intervals (various ones) and found 
recurring patterms.  For instance, sampling every timer tick (about 
55ms) produced a repeating pattern where every 16th sample had a 
delta well over 1200 while all the others were under 900, usually 
under 800.  Every so often the 16th sample delta was very low instead 
(between 400 and 600).

Plotting the samples in a noise sphere showed very definite streaks 
with a few spots in isolated areas.

So this is NOT a good method to use after all.

I'm still curious as to what causes the variation.  Maybe has to do 
with the CPU cache or port interfaces with the CMOS timer?

Rob. 

---
Send a blank message with the subject "send pgp-key" (not in
quotes) to <WlkngOwl@unix.asb.com> for a copy of my PGP key.