On Tue, 2004-03-30 at 09:37, Gerrit Voss wrote:
>
> well yes and no, for frame times most of the normal timers are ok, and
> at the time I did most of my fine grained timings I used Windows, that's
> why there the performaceCounter is implemented. Sometimes the problem is
> that this code is so small that you are tempt to use it without ever
> integrating it properly. Otherwise you often revert to profilers before
> actually adding your own timing code.
Yup. Unfortunately, high-resolution timers are very system-specific,
thus we haven't gotten around to implementing them.
> > BTW. Isn't the StatisticForeground doing some measurements as well?
> > How does this work?
>
> I guess just the normal system time.
Yup. Obviously we do have some time handling (in OSGTime.h), which is
good enough for many things, but it's not a high-res timer.
The perf counter would be a good enough solution for Windows, maybe you
can start with that (I didn't know we had that ;). In the long run it
would be nicer to have a library that wraps cycle counter for all the
relevant systems.
Dirk
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
