> Have you tried using INTEL_ONDEMAND_CLOCKMOD?  It currently doesn't do
> anything automatically, but it might help, or you could write a
> monitoring script.

nope.
I'm gonna try a ALL kernel then.

Will "[ while true ]; do envstat -r | tail -1; sleep 30; done" be enough ?
Or will I need something more kernel oriented ?

Regards,
  Jo