Mark Kettenis <mark.kettenis@xs4all.nl> wrote:

> So my take on this is: not now!  I'm in the same area to improve the
> suspend-to-idle code that was recently committed.

I agree with "not now".

> And while the idea of turning off cores that aren't used, in the end
> this depends on the scheduler to give you the right information.  And
> there are plans to make some drastic changes in that area as well.

I think the scheduler is very weak in making these decisions.  claudio
and mpi need to finish some of their work which simplifies the scheduler
first, then it will be more clear how to decide the actual question
which is

       "less compute resources needed"

I disabled the stupid algorithm a few years ago because (on newer machines),
it made it impossible to build a src tree on battery (because newer machines
have huge performance deltas).

A new algorithm with the same flaw is not desirable.

A knob to flip modes?  No, don't want that either.