[ntp:questions] 500ppm - is it too small?

Richard B. Gilbert rgilbert88 at comcast.net
Thu Nov 12 18:34:21 UTC 2009

Uwe Klein wrote:
> Richard B. Gilbert wrote:
>> Do we have any data on just how large a percentage of computer clocks 
>> are bad enough to exceed the 500 PPM limit?
> enough to kick off this  discussion. repeatedly.
>> And is there any reason for NTPD to stand on its head to cope with a 
>> clock that is broken that badly?
> you are lacking semantic finesse and differentiation ;-)
> A. you have crystals that behave eratic ( I would tag that "malfunction" )
> B. you have crystals that are off spec frequencywise. ( "off spec" )
> C. you have boards that lower apparent frequency due to spread spectrum 
> features
> D. you have boards that loose or gain ticks ("hardware")
> E. you have OSes   that loose or gain ticks ("OS bug")
> F. ....
> A is unfixable -> Broken
> B is fixable   -> could work if ntp would "yield"
> C is fixable   -> could work ""         ...
>     ( not 100% sure, depends on how spread spectrum is done in each 
> specific case )
> D is unfixable -> Broken , get other hardware
> E is fixable   -> but not via ntp which is OK
> There is an interesting short from Larry Niven's Draco Tavern series that
> takes up this semantic issue : "Grammar Lesson"
>> If you really believe that it's an arbitrary limit perhaps you should 
>> change it to 5,000 or 50,000 in your copy of NTPD and run it for a 
>> while.  Please let us know how well, or how badly, it works.
> I currently don't have a MOBO that would require such fiddling.
> uwe

Nor do I.  I see no point to expending the effort that would be required 
to make NTPD work on badly broken hardware.  If you do, you are welcome 
to do it yourself.  If you learn anything interesting or generally 
useful, please let us know.

I suspect that anyone who has a computer whose clock is off by 5,000 PPM 
would gladly donate it to the project.  If  you decide to undertake the 
project, feel free to request hardware donations!

More information about the questions mailing list