So, computer to computer sync is .1ms As the other two computers (the example and truetime.localn) both sync via gps you can't say they are off, cause they use an external source to sync, than the network.
Quoting John Pettitt <[EMAIL PROTECTED]>: > Patrick Domack wrote: >> Plus, I dunno if anyone has done stats on this, but I haven't been >> able to sync two ntpd servers closer than .2ms over gigabit >> switch. I won't say I have a good benchmark setup for this though. >> >> > remote efid st t when poll reach delay offset jitter > ===================================================================== > +GPS_NMEA(2) .GPS. 0 l 9 16 377 0.000 0.004 0.004 > oPPS(1) .TRUE. 0 l 13 16 377 0.000 0.011 0.002 > +truetime.localn .GPS. 1 u 10 16 377 0.391 -0.044 0.018 > skipjack.no-suc 69.25. 2 u 17 16 42 15.207 -0.104 0.313 > > This box is 0.044 ms off from another stratum one via a gigabit and > more interestingly 0.104 ms from a peer on the other end of a DSL line. > The secret to very close sync is very short poll times - something you > can do if you own both boxes as I do in this example. > > As an aside the box in question has three GPS refs - a GPS18, a PPS > from a Truetime GPS-DC and a Truetime GPS-100 stratum one neighbor. > > John _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
