[ntp:questions] NTP vs chrony comparison (Was: oscillations in ntp clock synchronization)

Unruh unruh-spam at physics.ubc.ca
Sun Jan 27 17:54:16 UTC 2008


brian.utterback at gmail.com writes:

>On Jan 25, 12:53=A0pm, Unruh <unruh-s... at physics.ubc.ca> wrote:
>> The Allan intercept is predicated on a very specific model of the noise in=

>> a clock ( as I recall basically random gaussian noise at high frequencies,=

>> and 1/f noise at low). It is not at all clear that real computers comply
>> with that.

>Then you should read Dr. Mills papers. You are correct that it is not
>necessarily
>the case that they comply, but Dr. Mills has demonstrated it
>experimentally many
>times for many different setups. Additionally, the behavior model is

And I have demonstrated it experimentally on my few computers that it does
not.-- step transients in the frequency are not part of the model. Now, I
suspect that if I took an unused computer and looked at the noise on it, it
would comply. 1/f noise is pretty universal at low f (and the reason is
poorly understood.) The question is what happens if the computer is in use-- used
heavily during the day ( higher CP at U temp) and not at night, and not at
weekends. Lots of power at 10^-5 Hz and harmonics, and .7 10^-8Hz.-- more
than would be predicted by 1/f

>nto actually
>the same over all possible configurations, which lead to his
>abandoning the FLL
>in the kernel code, since while theoretically useful it added needless
>complexity
>over the observed real world ranges.

>Pleasem Dave is not always right about everything, and can be
>convinced with data
>when he is wrong. But it is very hard to convince him with A Priori
>reasoning, since
>he has been doing that same reasoning for a lot longer than the rest
>of us, and has
>tested all his hypotheses on real systems.

Fair enough.

>Brian Utterback





More information about the questions mailing list