Re: [RFC PATCH v3 0/3] sched: simplify the select_task_rq_fair()

[Michael Wang <wangyun@linux.vnet.ibm.com>]

On 02/21/2013 02:11 PM, Mike Galbraith wrote:
> On Thu, 2013-02-21 at 12:51 +0800, Michael Wang wrote: 
>> On 02/20/2013 06:49 PM, Ingo Molnar wrote:
>> [snip]
[snip]
>>
>> 	if wake_affine()
>> 		new_cpu = select_idle_sibling(curr_cpu)
>> 	else
>> 		new_cpu = select_idle_sibling(prev_cpu)
>>
>> 	return new_cpu
>>
>> Actually that doesn't make sense.
>>
>> I think wake_affine() is trying to check whether move a task from
>> prev_cpu to curr_cpu will break the balance in affine_sd or not, but why
>> won't break balance means curr_cpu is better than prev_cpu for searching
>> the idle cpu?
> 
> You could argue that it's impossible to break balance by moving any task
> to any idle cpu, but that would mean bouncing tasks cross node on every
> wakeup is fine, which it isn't.

I don't get it... could you please give me more detail on how
wake_affine() related with bouncing?

> 
>> So the new logical in this patch set is:
>>
>> 	new_cpu = select_idle_sibling(prev_cpu)
>> 	if idle_cpu(new_cpu)
>> 		return new_cpu
> 
> So you tilted the scales in favor of leaving tasks in their current
> package, which should benefit large footprint tasks, but should also
> penalize light communicating tasks.

Yes, I'd prefer to wakeup the task on a cpu which:
1. idle
2. close to prev_cpu

So if both curr_cpu and prev_cpu have idle cpu in their topology, which
one is better? that depends on how task benefit from cache and the
balance situation, whatever, I don't think the benefit worth the high
cost of wake_affine() in most cases...

Regards,
Michael Wang

> 
> I suspect that much of the pgbench improvement comes from the preemption
> mitigation from keeping 1:N load maximally spread, which is the perfect
> thing to do with such loads.  In all the testing I ever did with it in
> 1:N mode, preemption dominated performance numbers.  Keep server away
> from clients, it has fewer fair competition worries, can consume more
> CPU preemption free, pushing the load collapse point strongly upward.
> 
> -Mike
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>