On 12/08/08 00:35, Li, Aubrey wrote:
> I just want to update the current working status of the c-state part of PAD.
>
> I think before put back into ON, we have to finish the following things:
>
> 1) cmt_balance() should be aware of the requirement of the current system,
> not messed power and perf up. That is, if perf is prefered, power related pg
> should be ignored.
>   

The idle PG should also be ignored when C-states are disabled.

> 2) [bug #4581] power idle group should be ahead of power active group for cmt 
> coalesce policy
>   
>> I failed to understand why it is marked as a duplicate of bug 4616
>>     

This does not seem right.

>
> 3) [bug #4616] cmt_balance schedules thread on any idle CPU instead of 
> shallow idle CPUs
>   

This fix helps perf/power in the lower ranges, but it hurts higher
load performance.  A good fix for next item #4 [bug 5444] will
help solve the high load issue.


> 4) [bug #5444] the system should not enter deep cstate at all when under 
> heavy load.
>   

> 5) Per current implementation
> =====================================================
>       pg = GROUP_ACCESS(cmt_pgs, level);
>
>             siblings = pg->cmt_siblings;
>             nsiblings = GROUP_SIZE(siblings);       /* self inclusive */
>             if (nsiblings == 1)
>                     continue;       /* nobody to balance against */
> =====================================================
> cmt balance will ignore a pg if its siblings size = 1. I found on pad head, 
> power idle pg's sibling
> size = 1, that means power idle pg will be ignored forever. \
> This problem is simiar as "[Bug 4015] Integer Pipeline processor group is 
> broken", which is
> closed and move the issue to 4616.
>   

Is there a bug for this?

> Is there anything I missed?
>   

4607 is not required for initial putback.

Eric and I are kicking around ideas for power scheduling
with Hyper Thread awareness.  That can probably be done
after initial putback?  (Fixing item #1 above will make this
a lower priority.)

-Bill

> Thanks,
> -Aubrey
>   


Reply via email to