On Fri, 2013-02-22 at 17:10 +0800, Michael Wang wrote: > On 02/22/2013 04:21 PM, Peter Zijlstra wrote: > > On Fri, 2013-02-22 at 10:36 +0800, Michael Wang wrote: > >> According to my understanding, in the old world, wake_affine() will > >> only > >> be used if curr_cpu and prev_cpu share cache, which means they are in > >> one package, whatever search in llc sd of curr_cpu or prev_cpu, we > >> won't > >> have the chance to spread the task out of that package. > > > > Nah, look at where SD_WAKE_AFFINE is set. Only 'remote/big' NUMA domains > > don't have it set, but 'small' NUMA systems will have it set over the > > entire domain tree. > > Oh, I missed that point... > > But I don't get the reason to make NUMA level affine, cpus in different > nodes share cache? doesn't make sense...
Contrary, it makes more sense, the more expensive it is to run 'remote' the better it is to pull 'related' tasks together. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
