I am using jetpack macrobenchmark library to measure UI performance(TTID, TTFD) and I got a situation when on the same device some benchmarks are different at different times, I understand that the device can have different performance at different moment of time, and sure each test runs 25-30 minutes, that also can make device to get hot and throttle(imho). I stared to research how to make device run always with the same performance and I'v got an information about CPU clock locking script, but I also got an note from official documentation of benchmarks which says that lockClock.sh is only necessary for microbenchmarks
Note: The lock clocking is only necessary for microbenchmarks.
Does that mean that with macrobanchmarking we'll not see difference whether we're using or we're not using clocking ?