Messages in this thread | | | Date | Mon, 29 Jul 2013 09:04:35 -0700 | From | Arjan van de Ven <> | Subject | Re: RFC: revert request for cpuidle patches e11538d1 and 69a37bea |
| |
On 7/29/2013 9:01 AM, Lorenzo Pieralisi wrote: > >> Coupled C states on this level are a PAIN in many ways, and tend to totally suck for power >> due to this and the general "too much is active" reasons. > > I think the trend is moving towards core gating, which resembles a lot to what > x86 world does today. Still, the interaction between menu governor and > cluster states has to be characterized and that's we are doing at the > moment.
I suspect that we (Linux) need a different governor than menu for the coupled case. Fundamentally you're going to need different algorithms that are much closer tied to the behavior of the specific hardware; the current menu governor uses an abstract and simple hardware model (fully independent). For specific hardware that deviates from the model, something different is clearly needed.
It's an open question for me if just the predictor is the part that you split out, or if the whole governor needs to be split out. (but frankly, if you take the predictor out of the menu governor, not a whole lot is left)
| |