Stefan Ott wrote:
Hi!
I have an IMHO rather weird behaviour: If I stop xfce4-panel, I get (as
expected) way less wakeups (ca. 65 instead of 110). However, instead of
spending more time in C4, this is how the CPU's state changes:
With the panel:
Cn Avg residency (10s) Long term residency avg
C0 (cpu running) ( 3.8%)
C1 0.0ms ( 0.0%) 0.0ms
C2 9.2ms (53.9%) 0.0ms
C3 6.2ms (25.8%) 0.0ms
C4 5.9ms (16.4%) 4.9ms
Without the panel
Cn Avg residency (10s) Long term residency avg
C0 (cpu running) ( 1.0%)
C1 0.0ms ( 0.0%) 0.0ms
C2 17.0ms (74.2%) 0.0ms
C3 13.7ms (24.6%) 0.0ms
C4 5.2ms ( 0.2%) 4.8ms
As you see, C4 is hardly used anymore. How does this make sense?
this is very likely the known bug in the C-state code of
2.6.21/2.6.22... the fix exists as a patch at this point (by rewriting
the whole lot ;-)
_______________________________________________
Power mailing list
[email protected]
http://www.bughost.org/mailman/listinfo/power
