[ntp:questions] Re: Unusually accurate PC clock? 3.326 ppm drift? Is this possible?

Terje Mathisen terje.mathisen at hda.hydro.com
Fri Jan 28 21:04:30 UTC 2005

elickd at one.net wrote:

> In testing our systems, I started the xntpd daemon on this particular
> IBM Netfinity 5500 running SCO 5.0.5.  Xntpd always stays synced to the
> three upstream servers it's pointed at (I know it should have more,
> it's only a test) and the drift number is consistantly around 3.
> root at xxxxxx / # cat /etc/ntp.drift
> 3.326
> This machine has always been rock solid, but I'm having a hard time
> believing such an increadible drift value isn't the product of a bug or
> misconfiguration of xntpd.
> Have any of you ever seen such a stable, unmodified PC clock?


This is just the 'luck of the draw', and the actual value doesn't really 
matter at all:

What's important is the second-order effects, i.e. how much is this 
value changing over the day/year?


- <Terje.Mathisen at hda.hydro.com>
"almost all programming can be viewed as an exercise in caching"

More information about the questions mailing list