I am going through Macrobenchmarking documentation : https://developer.android.com/topic/performance/benchmarking/benchmarking-overview
It mentions:
It allows you to control compiling, starting and stopping your application to directly measure actual app startup or scrolling, instead of only the specific jitted, hot functions, with all disk accesses cached.
I understand the first part about controlling compile, startup as benchmark provides us with those options while writing the tests. But unable to understand what the second part means only the specific jitted, hot functions, with all disk accesses cached.
Can someone help me understand this. Is this related to AOT compilation or something else?
With the macrobenchmark library you gain full control over several states in contrast to the Microbenchmark library.
This is related to JIT (Just-in-Time) and AOT (Ahead-of-Time) compilation.
I agree that the phrasing is targeted at developers who already have an understanding of these libraries and it will be updated shortly.