Richard B. Gilbert wrote: > David J Taylor wrote: >> Richard B. Gilbert wrote: >>> David J Taylor wrote: >>>> David Woolley wrote: >>>> [] >>>>> You won't get millisecond accuracy on Windows. Although the >>>>> software clock can be disciplined to better than a millisecond, >>>>> applications can only read to one tick, which is 10ms by default >>>>> and 1ms with the fastest multi-media timers (which risks lost >>>>> ticks). >>>> >>>> David, do you have a reference for your "risks lost ticks" statement >>>> as applied to Windows? >>>> >>>> Thanks, >>>> David >>> >>> ISTR that there have been previous references in this newsgroup. >>> Specifically, EIDE disk drivers can mask or disable interrupts for a >>> period long enough to cause a "lost tick". I believe the group is >>> archived somewhere. . . . >> >> Thanks, Richard. I can imagine that any disks working in PIO mode >> wouldn't be welcome friends either. The last EIDE desktop PC I bought >> was some seven years ago, so I hope it's a problem which won't affect >> too many people, and all being well not those starting from scratch. >> > > I don't know if SATA is any better. I don't run ntp on my PC.
I don't think the Windows MM timer problem which is worked around by the -M switch is due to lost timer ticks. If it was then you would always observer a step in the same direction. However, if you don't use -M then the interpolated system time seems to step in one direction when another program sets the MM timer to highest resolution, and steps back by the same amount if that application terminates. So this just seems to be a problem in the way the Windows kernel converts timer ticks at different rates to system time. Martin -- Martin Burnicki Meinberg Funkuhren Bad Pyrmont Germany _______________________________________________ questions mailing list [email protected] https://lists.ntp.org/mailman/listinfo/questions
