Messages in this thread | | | Date | Wed, 8 May 2013 13:15:16 +0200 | From | Peter Zijlstra <> | Subject | Re: [PATCH v5 3/7] sched: set initial value of runnable avg for new forked task |
| |
On Tue, May 07, 2013 at 11:24:52AM +0800, Alex Shi wrote: > It will give new forked task 1 ms extra running time. That will bring > incorrect info if the new forked goes to sleep a while. > But this info should benefit to some benchmarks like aim7, > pthread_cond_broadcast. So I am convinced. :) > > What's your opinion of this, Peter?
We're going to get it wrong anyhow (lack of crystal ball instruction); overestimating the load seems like the best deal wrt latency for new tasks.
The only other option we could try (not sure its even worth trying) would be half load.
| |