[ntp:questions] Re: linux system clock egregiously slow!

Hal Murray hmurray at suespammers.org
Thu Jan 22 07:32:23 UTC 2004


>It is actually 149605480.  I haven't done the math in a while, but
>taking our actual CPU clock frequency into account it is within a few
>PPM of two instruction cycles per counter increment.  Whereas the
>linux system clock is closer to 800PPM slow!

(At least) One of us can't do math.

You are expecting 150.0  You get 149.6

That's off by
    ~3 parts per 1500 or
     2 per 1000 or
     2000 per million.

What did I fatfinger?

Take the box apart and look around for the crystals.  Can you
find a magic clock generator chip near the CPU.  A crystal will
connect directly to it.  14.?? MHz crystals are popular.

It will have 30-48 pins.  Big for anything but a major chip.
If you find something suspicious, use google to find the
data sheet.  You are looking for a chip that makes all the clocks
for the system.  300 MHz for your CPU, something for PCI, something
for the memory system...  (The CPU chip may contain it's own PLL.)

2000 ppm is 0.2%.  That's the average frequency deviation.  If it's
sine wave modulation, the peak is twice that.  That's the right ballpark
for the anti-EMI spread spectrum clock generation chips.



-- 
The suespammers.org mail server is located in California.  So are all my
other mailboxes.  Please do not send unsolicited bulk e-mail or unsolicited
commercial e-mail to my suespammers.org address or any of my other addresses.
These are my opinions, not necessarily my employer's.  I hate spam.




More information about the questions mailing list