[ntp:questions] Dispaying nanoseconds

Eugen COCA ecoca at eed.usv.ro
Wed Feb 14 20:23:49 UTC 2007

On Feb 14, 7:05 pm, "David L. Mills" <m... at udel.edu> wrote:
> Eugen,
> The displays in ntpq are in milliseconds and parts-per-million. The
> offset display is in fractions of milliseconds with three digits of
> significance, which resolves the nanoseconds. Is there something more
> you need?
> Dave


ntpq displays microseconds only:

ntpq> pe
     remote           refid      st t when poll reach   delay
offset  jitter
+GPS_NMEA(1)     .GPS.            0 l   58   64  377    0.000
0.000   0.004

I understood the offset and the jitter are in miliseconds. Is the PPS
source will be very precise, the values for offset and jitter
displayed by ntpq will be 0.000 miliseconds.

With ntptime I obtain something like this:

ntp_gettime() returns code 0 (OK)
  time c97debd1.d6c907cc  Wed, Feb 14 2007 22:13:05.839, (.839005942),
  maximum error 6466 us, estimated error 1 us, TAI offset 0
ntp_adjtime() returns code 0 (OK)
  modes 0x0 (),
  offset -0.180 us, frequency -16.215 ppm, interval 256 s,
  maximum error 6466 us, estimated error 1 us,
  time constant 4, precision 0.001 us, tolerance 496 ppm,
  pps frequency -16.215 ppm, stability 0.004 ppm, jitter 1.779 us,
  intervals 91977, jitter exceeded 594674, stability exceeded 36,
errors 230.

So, the kernel (?) knows the offset is 0.180us = 180 nanoseconds but
ntpq said it's 0.000 us (the rounded value for 0.18).

My ideea is if there is a possibility to display the offsets with 1E-9
precision - with ntpq or any other program who know how to extract the
data from ntpd (if it is stored inside with this precission).

I saw multiple graphs from professional public time services where the
offset between the time served by the computer and a reference clock
are displayed in nanoseconds. Are there any other technical means to
measure these offsets (like frequency counters or even counters) ?

Thank you !

More information about the questions mailing list