Gary E. Miller gem at rellim.com
Mon Nov 7 20:38:25 UTC 2016

Yo Achim!

On Mon, 07 Nov 2016 21:28:27 +0100
Achim Gratz <Stromeko at nexgo.de> wrote:

> >>  This data was for 86400 raw PPS
> >> timestamps (which you are seemingly not recording), you are talking
> >> the loopstats.  
> >
> > Correct.  I care about the end result.  But when I look attt my raw
> > data I actually see worse results in the unaveraged data.  
> You still mistake the loop offset for that end result you so vaguely
> talk about.

Vaguely?  No, I'll just let my data speak for itself:


When you have data that looks better, no matter how you get it, I'd
like to try to replicate.

> >> So in fact I do get better numbers than yours at least at the tail
> >> ends of the distribution.  
> >
> > Facts not in evidence.  
> You've conveniently snipped those numbers so you can make it look like
> none were provided.

None useful.  Seemed like cherry picking to me.  Please provide ntpviz URLs
so we can compare apples and apples.

> > Please post ntpviz graphs and stats.  
> I gave you the statistics you've asked for and are now ignoring.  I
> simply don't run ntpviz.

Very few stats, with no backup.  See above.

> > Yuu think you are seeing xtal aging in 24 hours?  I seriously doubt
> > it and you'll need a lot longer data collection to prove it.  
> You might want to read a whitepaper or book on how crystals age, there
> is a choice of more than a handful you can just download, from all the
> usual suspects.  Or maybe you don't want and keep your opinion.
> Either way is fine with me.

Been there, done that.  If you have one that shows 'reversible aging'
please send the link.

> What you're seeing here is most likely caused by stress relaxation
> (thermal hysteresis), since the long-term aging actually goes the
> other way.

Then it is not aging, is it?  I never said 'thermal hysteresis' was
not an effect.

But this is all arguing hwo many angles dance on the head of a pin, unless
and until you can get the end results of NTP (loopstats) looking better.

I have long thought the signal processing in ntpd was sub-optimal.  I
suspect there are changes in there that would help a lot, see if you can 
find one and share.  Then we ccan try to replicatet.

Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem at rellim.com  Tel:+1 541 382 8588
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 455 bytes
Desc: OpenPGP digital signature
URL: <http://lists.ntpsec.org/pipermail/devel/attachments/20161107/0fec646b/attachment.bin>

More information about the devel mailing list