I think I even commented on the bugreport that it would be best to do the frequency scaling based on interrupts from the voltage regulators on board, but that piping all the necessary drivers/events to do that with low enough latency to not cause an under-voltage reset was far more work than was feasible.
Thats also why 60fps video dropped frames at the start of the video if you were recording video with less than 60% battery.
Anyone care to dig out the relevant commit...?
My Android (1st Gen Moto G, running Lineage OS, about 3 years old) shuts down once the battery level hits 15%. You can actually see it change from 15% to 1% for one brief moment before it shuts down.
Sure, the battery is old, but this behavior is unlike the behavior of any other OS/device I've seen. I've seen some laptops with seriously messed up batteries and as far as I can remember, the OS reports a smooth discharge profile without steep drops and without and strange "surprises".
I think Android is probably not taking battery health into account when it reports remaining capacity, but that is only a guess.
It is a bit tricky, but well worth it.
https://www.makeuseof.com/wp-content/uploads/2012/02/power_s...
My Samsung Android phone shuts down at 5-7% battery life. And the battery dies much faster these days. But it's never "inaccurate." It doesn't jump from 50% to 2% over a course of five minutes.
Other than something else being wrong with your OS, my only guess there would be the battery got bad so quickly the OS hasn't adapted to the new life cycle yet by removing enough of the old datapoints.
100-60 is still decent.
I have a long history with almost every Android device. The device doesn't get slower, the battery life just gets terrible.
Other devices suffer from similar effects - not as bad, but most devices certanly lose quite a bit of I/O performance (especially writes) over time which makes OS run noticably slower.
There are literally billions of Android devices. Benchmarking file system performance is trivial. Ergo, where are the benchmarks supporting your broad claims? This should be incredibly easy to demonstrate. We do know that it happened pre TRIM because people had factory new and after use benchmarks, and could show the decline.
The biggest killer, though, is storage with the graphite electrode more than 50% lithiated - there is a gigantic step-up in capacity fade when batteries are stored with the State of Charge over around 90%, which Tesla are very smart to avoid.
BTW the car won't actually let you drive at low charge, so yes, you can be stranded because the car refuses to damage its battery. That's why it comes with a towing service.