[ntp:questions] A faster settling NTP

unruh unruh at wormhole.physics.ubc.ca
Fri Dec 25 06:34:08 UTC 2009

On 2009-12-25, David J Taylor <david-taylor at blueyonder.delete-this-bit.and-this-part.co.uk.invalid> wrote:
> "unruh" <unruh at wormhole.physics.ubc.ca> wrote in message 
> news:slrnhj7qd9.3p4.unruh at wormhole.physics.ubc.ca...
> []
>>> Why not take a different approach, then?  Start with chrony to make the
>>> system rapidly stable, and then turn control over to NTP.  Is something
>>> like that possible?
>> Why not just stick with chrony. But I thought we were discussing Windows
>> systems, on which chrony does not run. And if someone has rewritten
>> chrony to run on windows, there is no point in handing over to ntp. As I
>> said chrony is about 2 times better ( lower standard deviation) than
>> ntp. Why would you want to hand over control to a worse program?
> I had wondered whether that might be an acceptable solution for some folk. 
> My original question was not OS-specific.  As you say, until chrony is 
> available for a wider range of operating systems and supports GPS/PPS 
> reference clocks it is but of academic interest.  As NTP already provides 

It does (now) support GPS/PPS refclocks. 

> that infrastructure, enhancing the initial convergence rather than having 
> to re-invent the whole wheel seems a better approach to me.
> Happy Christmas all all our readers!
> David 

More information about the questions mailing list