[LinuxPPS] assert/clear

[Miroslav Lichvar]

On Thu, Jul 29, 2010 at 04:33:38PM +0200, Remco dB wrote:
> > 5-35ns offset is very impressive, is that the offset reported in
> > loopstats?
> 
> No, RRD-tool makes an average (with the AVERAGE option). 
> I reckon that this is representative enough. Look at: 
> http  remco.org/ntp/helium-ntp-offset.png

Ah, ok. Root mean square would be a better measure, but I'm not sure
if you can calculate it easily in rrdgraph.

> Bear in mind that the PLL-clock crystal (14.31818 MHz) of helium is thermally 
> stabilized.

Nice. What polling interval do you use? If the clock is very stable
using longer polling interval might improve the accuracy a bit.

-- 
Miroslav Lichvar