[ntp:questions] NTPD can take 10 hours to achieve stability

Dave Hart hart at ntp.org
Tue Apr 19 09:56:47 UTC 2011


On Tue, Apr 19, 2011 at 6:03 AM, unruh <unruh at wormhole.physics.ubc.ca> wrote:
> What is worse is that modern cpus in order to conserver energy keep
> changing their rate, which makes them really difficult to use as a good
> rate.

At least that behavior can usually be configured away by instructing
the system to run at a fixed rate regardless of load.  Much more
pernicious are systems which use spread-spectrum CPU clocking to dodge
RF emissions limits.  This loophole (intentional idiocy by complicit
regulators) is often not configurable at all, and can result in 1000
PPM or more of local clock wander.

Beware ntpd on newer laptops...

Cheers,
Dave Hart



More information about the questions mailing list