Date: Sat, 24 Jan 2004 20:45:14 -0700 (MST) From: "M. Warner Losh" <imp@bsdimp.com> To: dillon@apollo.backplane.com Cc: current@freebsd.org Subject: Re: DragonflyBSD kernel clock improvements Message-ID: <20040124.204514.109171577.imp@bsdimp.com> In-Reply-To: <200401242031.i0OKVD8A037265@apollo.backplane.com> References: <44827.1074974041@critter.freebsd.dk> <200401242031.i0OKVD8A037265@apollo.backplane.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message: <200401242031.i0OKVD8A037265@apollo.backplane.com>
Matthew Dillon <dillon@apollo.backplane.com> writes:
: No apic timer. No acpi timer. No TSC garbage. none of that.
We've found at work that while the 8254's timer tends to be the best
universally available time keeping device, it is only as good as the
quarts oscillator driving it. Other time keeping can be more precise.
We've found that systems with APM disabled have a more stable NTP
server when they are running on CPUs with TSC using the TSC
timecounter, but about a factor of 5, mostly due to a reduced jitter
in timestamping the PPS interrupts that we generate. Of course, we
don't care one whit about SMP support for these systems.
Maybe there are better ways to obtain these results. While not
directly releated to the nanosleep work, but just another point of
view.
Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040124.204514.109171577.imp>