Re: sched: Consequences of integrating the Per Entity Load TrackingMetric into the Load Balancer
From: Mike Galbraith
Date: Fri Jan 04 2013 - 06:41:15 EST
On Thu, 2013-01-03 at 16:08 +0530, Preeti U Murthy wrote:
> Hi Mike,
> Thank you very much for your feedback.Considering your suggestions,I have posted out a
> proposed solution to prevent select_idle_sibling() from becoming a disadvantage to normal
> load balancing,rather aiding it.
> **This patch is *without* the enablement of the per entity load tracking metric.**
> This is with an intention to correct the existing select_idle_sibling() mess before
> going ahead.
Well, on the bright side, this is much kinder to light load tbench on
4x10 core (+ht) box than current bounce happy select_idle_sibling(). At
the heavy load end I'm losing heaping truckloads of throughput though,
likely some migration rate limiting would fix that up.
select_idle_sibling() doesn't have much of a pretty face these days, but
its evil hag face shows more prominently than ever, so making it dead is
a great goal ;-)
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/