[Lse-tech] Re: NUMA scheduling

[Mike K. <kr...@us...>]

On Fri, Feb 22, 2002 at 07:59:02PM +0000, Alan Cox wrote:
> > - Topology discovery.  To create cpu sets, a routine numa_sched_init()
> >   is called after we know how many cpus there are in the system.
> >   Inside numa_sched_init(), there are hard coded values used to indicate
> >   the number of cpus per set, and the 'distance' between sets.
> >   Obviously, creation of cpu sets would be based on topology information
> >   that is not available at this time.  In addition, there should be
> >   support for (or at least an understanding of) non-symmetric and
> >   multi-level topologies.
> 
> Would you expect that to be in user space eventually (I'm thinking about
> hot plug CPU's for one here) ?

Yes, I believe there should be interfaces for topology discovery both
in kernel and user space.  Some people on the lse-tech list have been
working on this.  Although, I have to admit that I haven't been
closely following those discussions.  Perhaps the people working on
this could let you know what they are planning.

-- 
Mike