RE: NSTimer fires ~0.0045 seconds early?
RE: NSTimer fires ~0.0045 seconds early?
- Subject: RE: NSTimer fires ~0.0045 seconds early?
- From: Christian Mike <email@hidden>
- Date: Mon, 6 Aug 2001 08:17:33 -0500
This takes me back to my DEC PDP-11 days. (I'm not THAT old, by I'm aging.
:-) ). The granularity was in "ticks", which where defined to be the AC line
frequency (60 ticks per second). In reality, this turned out to be the
number of AC line crossings.
The implication here is that if you scheduled an event to happen "one tick"
later, it happened at the next AC line crossing, which could be from almost
zero to 16.67 milliseconds later. So if you needed a delay of at least "x"
ticks, you always made the delay to be "x+1".
I'm guessing that something along this order is happening here. I know that
in the earlier days of IBM compatible stuff (my age keeps showing!!!), the
clock quantization was 32.5 milliseconds. I haven't done enough digging into
the Mac OS X or current technology PPC documentation to know what happens
today.
Michael Christian
Thomson multimedia Inc.
-----Original Message-----