cx_lowest and CPU usage

Andriy Gapon avg at icyb.net.ua
Thu Feb 14 13:35:57 PST 2008


on 14/02/2008 23:19 Kevin Oberman said the following:
> 
> It certainly blocks ever reaching C3 because of the continual activity
> it generates, but I can see that it might also increase the interrupt
> time while in C2. That said, I just did some testing on my laptop in C2
> and, while I see a small increase in interrupt load, it is not anywhere
> near what is being reported in this thread.

Kevin,

a lot also depends on how idle/interrupt/user/etc is calculated.
See my recent post to -hackers.
When I had only an intuitive feeling of how those things are calculated
I also suspected that some extra calls, transition times, etc might be
to blame. But after I briefly looked through the code my opinion was
somewhat altered.

-- 
Andriy Gapon


More information about the freebsd-acpi mailing list