On Sun, Nov 16, 2003 at 02:56:22AM +0900, Izumi Tsutsui wrote:
> This means crystal clock used on my U5 has actually bad accuracy?

Is that running -current? All my sparc64 machines have pretty good ntp
statistics, since we switched over to using the MI microtime code based
on cycle counters (%tick).

> Or "using %tick at 269MHz as system clock" is the problem?

I doubt that. It is a pretty exact method, if the firmware does not lie about
the frequency (and there are no software bugs).

Martin