[avr-gcc-list] AVR Studio and Hardware: results don't match. Any clue?

[Leonardo B. Oliveira]

Hello everyone,

I've measured running time for a program in AVR Studio and also in MICAz sensor node.
MICAz is featured w/ the ATmega128 microcontroler.  However, the results I've got don't match. In reality, the results in the simulator were much faster. I know that when loading software on nodes other software is also loaded (
e.g. TinyOS). But, at least, they are not supposed to interfere that much in the performance. Besides, the results I've got in the simulator are twice as fast.

So my questions are
1) Is AVR Studio an accurate simulator?
2) I also have the STK300 AVR JTAG ICE. If I run the same program using  both the board and the AVR Studio (by means of the JTAG), will the clock cycles exhibit in the simulator be taken from the real board? Or will they be the same of running the simulator alone?

Anyway, I'm newbie concerning AVR platforms and I'd like to know how performance numbers are usually measured in those devices. Any help would be much appreciated.

Thanks in advance.

-- 
Cheers,

  Leonardo

Alice:..wondering u could help me find my way
Cat:It depends on where u want to get to
Alice: It really doesn't matter, as long as...
Cat: So it really doesn't matter which way u go