On 2015-01-13, Stuart Henderson <s...@spacehopper.org> wrote: > 1x e-03 (macppc mac mini)
Thanks to all who replied. Macs and Suns seem to be particularly bad. I counted zeroes again and just realized that the combo of ntp.org and ntp_adjtime() can only correct offsets up to 5e-4 (500 ppm). I wonder if our timekeeping on these archs is broken. > 2x e-08 (esxi) Oooh, interesting. I hadn't considered VMs that actually keep time. Anyway, I was asking because we want, for compatibility reasons, to switch OpenNTPD's drift file format from an unscaled frequency offset to ppm, and it would be nice if we could just tell from the magnitude of the number what unit it's in. See the patch on tech@. -- Christian "naddy" Weisgerber na...@mips.inka.de
