[ntp:questions] NTP Performance on WINNT

Unruh unruh-spam at physics.ubc.ca
Sat Jul 25 17:48:12 UTC 2009


"David J Taylor" <david-taylor at blueyonder.not-this-part.nor-this.co.uk.invalid> writes:

>paul wrote:
>[]
>> Thanks, but I mean something which do not rely on the output of
>> ntpq.exe.

You can get say a gps with PPS and attach it to say the parallel port of
your machine, writing and intrrupt service routing to grab the interrupt
times. Then look at those to see how far off your system is. Those times
should be good to a usec or so. You could of course then use ntp to use
that data to discipline the local clock. In that case the results would
show up in ntpq. 
Also, instead of using ntpq, you could use the output in
/var/log/ntp/peerstats to see directly the offset on each measurement.



>My NTP Monitor uses NTP network calls to determine the offset of the PCs - 
>it doesn't use ntpq.  However, in that sense, it does rely on the output 
>from NTP.

>What problem are you trying to solve, and how accurately do you need to 
>measure?

>Cheers,
>David 




More information about the questions mailing list