Re: [PATCH 3/4] sched,numa: implement numa node level wake_affine

2017-06-26 Thread Peter Zijlstra
On Fri, Jun 23, 2017 at 12:55:29PM -0400, r...@redhat.com wrote: > From: Rik van Riel > > Since select_idle_sibling can place a task anywhere on a socket, > comparing loads between individual CPU cores makes no real sense > for deciding whether to do an affine wakeup across

Re: [PATCH 3/4] sched,numa: implement numa node level wake_affine

2017-06-26 Thread Peter Zijlstra
On Fri, Jun 23, 2017 at 12:55:29PM -0400, r...@redhat.com wrote: > From: Rik van Riel > > Since select_idle_sibling can place a task anywhere on a socket, > comparing loads between individual CPU cores makes no real sense > for deciding whether to do an affine wakeup across sockets, either. > >

[PATCH 3/4] sched,numa: implement numa node level wake_affine

2017-06-23 Thread riel
From: Rik van Riel Since select_idle_sibling can place a task anywhere on a socket, comparing loads between individual CPU cores makes no real sense for deciding whether to do an affine wakeup across sockets, either. Instead, compare the load between the sockets in a similar

[PATCH 3/4] sched,numa: implement numa node level wake_affine

2017-06-23 Thread riel
From: Rik van Riel Since select_idle_sibling can place a task anywhere on a socket, comparing loads between individual CPU cores makes no real sense for deciding whether to do an affine wakeup across sockets, either. Instead, compare the load between the sockets in a similar way the load