Messages in this thread | | | From | "Tian, Kevin" <> | Date | Fri, 17 Oct 2008 10:43:24 +0800 | Subject | RE: [linux-pm] [RFC PATCH 0/4] timers: Timer migration |
| |
>From: Arjan van de Ven >Sent: Thursday, October 16, 2008 10:38 PM >On Thu, 16 Oct 2008 17:12:39 +0530 >> >> >> Here's a brief introduction as to why we need timer migration. >> > >as I've said before, for reducing power consumption I don't buy into >needing this concept.... > >If you have high frequency timers, you've lost no matter what (and we >fixed almost all of them in the ecosystem); if you don't have those >then this patch doesn't really add anything. >(and the range-hrtimers fixes the intermediate state) >
Does range-hrtimers work in per-cpu style or globally? If the former I guess this timer migration may help range-hrtimers feature to further align rest timers. If globally, then small room remains.
Arun, did you have any actual power improvement data with it?
Thanks, Kevin
| |