[time-nuts] Win XP and NIST time

Dan Kemppainen dan at irtelemetrics.com
Tue Mar 26 14:21:39 EDT 2013


On 3/26/2013 2:43 AM, time-nuts-request at febo.com wrote:
>> >albertson.chris at gmail.com  said:
>>> >>I think you can get Windows to run at the "few milliseconds" of error range
>>> >>with the standard NTP distribution.
>> >I don't think I've seen anything that bad, but it's easy to be off by 100s of
>> >ms if I download something big like a CD or a long video
> What I meant was that "you can get Windows to run at the "few
> milliseconds" of error range".  not that it will be that good in every
> case.   What I really meant was "don't expect uSec level timing from a
> normal Windows system".   But is CAN be as good as a few mS.
>
> In your case you'd need GPS or some other reference clock to get
> there.  Most people are getting no better than 10s of mS over the
> Internet
> -- 

Oddly enough, I had a strange way to compare two windows system today. I 
had a webex meeting, and the other party opened their clock and I could 
see seconds ticking away. I opened my clock, and the seconds were only 
about a second apart, mostly due to latency in getting data across the 
network. This isn't the first time I've done this.

This is, in reality impressive, considering the delays in moving video 
data across the network. So, two windows boxes half way across the US 
showed the same time to within network latency of around a second or so. 
(TZ offset ignored, of course)

Keep in mind, we are after all, taking about windows. An operating 
system that IS NOT real time operating system. (You think it is, try 
move a continuous stream of a few 6+ MBytes/Sec data to it!)

As much as this gets to some time nuts (you know who you are! :) ), on 
windows, this is good 'enuf' for me!  (Even when measuring how long it 
takes to calculate PI to 80e9 places...)


Dan



More information about the time-nuts mailing list