>> Note that on mica motes, CounterMicro32C stops counting while the mote >> is asleep (otherwise the mote wouldn't be able to sleep ;-)). Also >> it's not exactly counting microseconds (mica's run at 7.XX MHz, so the >> time unit for CounterMicro32C is 7.XX/8 us). >> >> David Gay ... > I think the issue about sleeping will not be problematical in my case, > because i will get the time from counter before calling the method and > after finishing the method, therefore i think in this period the mote > will not sleep. Or am I thinking wrong ? (Sorry, I am new in programming > with tinyOS) > > For the exact counting, I think micaz runs at 8Mhz. Should I also > calculate 7.37/8 to get exact timinig results ? > > Osman Ugus
One thing to think about, do you want: 1. absolute clock execution 2. processor time 3. a relative metric of code efficiency 4. or something else? Approximating runtime is good for most questions, unless you need to compare across platforms, then transform it later. If it runs across sleep boundaries, do you want to time that or not? Is it important? If you want 3, then it certainly is not, and could throw your statistics off. Just trying to help out, -Ben _______________________________________________ Tinyos-help mailing list [email protected] https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
