> On Fri, Jul 10, 1998 at 10:22:26AM -0400, Richard B. Johnson wrote:
>
> > I have by good authority that the "next generation" of PC Mother-boards
> > will not use quartz crystals to generate clocks. They will use ceramic
> > resonators such as the kind made by muRata (www.murata.com). These
> > devices are surface-mount "chips" to reduce cost, and provide only
> > 2 percent accuracy as compared to 0.001 percent accuracy from typical
> > crystals presently used.
> >
> > Since time-keeping is important, and inaccuracies are cumulative, we
> > will have to come up with software fixes, possibly using the CMOS timer
> > for periodic re-calibration.
>
> These machines are loosing boxes anyway. For example the xntpd software
> PLL will most probably loose synchronisation every couple of minutes,
> if it syncs at all.
>
> For certain applications such a type of board is plain unuseable.
>
> Ralf
>
But.... they're gonna be called 430MHz Pentiums. It's a shame to throw
away that kind of power because the machine can't keep time! The CMOS
chip will always keep time 'good enough'. Note that 'better' is the
enemy of 'good enough'.
We need to use timer-channel 0, connected to IRQ0 for context-switches
because it's the highest priority interrupt. However, we could use the
CMOS timer, which is connected to a lower-priority interrupt, for the
jiffy-counter and basic time functions. Or the PLL attempt, really
a frequency-lock, already implemented, could sync to the CMOS timer.
This is something that should be planned because soon only 'obsolete'
boxes will keep time.
Cheers,
Dick Johnson
***** FILE SYSTEM MODIFIED *****
Penguin : Linux version 2.1.108 on an i586 machine (66.15 BogoMips).
Warning : It's hard to remain at the trailing edge of technology.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html