The hackers working on hi-resolution ntpd clock development, for example, use a PPS output from a GPS or other "stratum 0" clock and compare this at the microsecond level with the disciplined PC clock.
Interrupt latency matters a lot if you want PPS interfaced to a PC for useful results at the nanosecond level but some do it. Ntpd has excellent support for PPS but PPS exists outside it too. There are other scientific and instrumentation time synchronization protocols that vary by the sub-field. The concepts of drift, latency, round-trip delay and jitter exist inside of ntpd and any decently developed time synchronization protocol. You will want to check out the time-nuts mailing list if you want to deal with time and frequency measurements more broadly. There are mailing lists and RFC's devoted to other time synchronization protocols. Many of them are quite primitive compared to the high state of development and implementation of ntpd. -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Mag Gam Sent: Tuesday, August 03, 2010 7:38 AM To: [email protected] Subject: [Pool] measuring time If tools like ntpq didn't exist how does one measure time? Lets say I am trying to compare X number of time protocols and I want to see which one is the closest to a stratum 1 or GPS clock, how should I go about measuring the skew? What is the scientific way of doing this? _______________________________________________ pool mailing list [email protected] http://lists.ntp.org/listinfo/pool _______________________________________________ pool mailing list [email protected] http://lists.ntp.org/listinfo/pool
