[ntp:questions] Sudden drop in frequency after software update

David J Taylor david-taylor at blueyonder.neither-this-bit.nor-this.co.uk
Mon Jan 19 07:47:04 UTC 2009


David Woolley wrote:
> David J Taylor wrote:
>
>>
>> Would it be any better for the routine to lie, and just give to the
>> nearest MHz, rounded down?
>
> That would result in an excessive error for those not using time
> synchronisation software,

I might say /tough/!

> and would produce extreme instability if the
> frequency was very close to the exact MHz.

Yes, that's why I suggested rounding, but there will undoubtedly be /some/ 
frequency for which it didn't work well.  Suggestion withdrawn.

Why do people mess with Linux like this?

Are current versions of FreeBSD much better for timekeeping than Linux?

David 




More information about the questions mailing list