4 ms·
You do if that's similar to the expected deployment environment.
by dalke 14y ago
You do if that's similar to the expected deployment environment.
- stefantalpalaru 14y agoOnly if it's something you can replicate exactly for each run of the benchmark.
- dalke 14y agoThere's no need for perfect repeatability when statistical analysis is good enough. After all, even if you have control down to the iron, the randomness in external interrupts and the effect of temperature on the hardware will cause some unpredictability.