Re: status of NetBSD SMP support

[Thor Lancelot Simon <tls%panix.com@localhost>]

On Mon, Aug 08, 2016 at 04:42:08PM +0000, David Holland wrote:
> On Mon, Aug 08, 2016 at 08:14:50AM +0530, Cherry G. Mathew wrote:
>  > > The cores are coming! The cores are coming!
>  > >
>  > > The next round of Intel Xeon chips have core-counts in tens. UltraSPARC
>  > > Niagara (T1, T2, et alia) have had lots of cores for quite some years.
>  > > Multicore ARMs are in all mid-to-high-end smartphones & tablet computers.
>  > >
>  > > Can the NetBSD kernel use them all efficiently & effectively?
>  > >
>  > > Have the user land applications which can be parallelized been
>  > > parallelized?
>  > >
>  > > What is the current status of NetBSD SMP system support? Which kernel
>  > > subsystems still need work, and which are done?
>  >
>  > Bit late reply, but we have had support for upto 256 CPUs for ages now.
>  > 
>  > I don't know much about userland though.
> 
> Last I remember anyone reporting hard results, the scaling worked to
> ~16 but not to ~32 and the uvm page queue lock was the chief culprit.
> Dunno what if anything's been done about that...

There's a fragmentary discussion of it from around 2010 in the mailing list
archives, but something must have been done as that particular limitation
seems to have gone away.  Our build cluster nodes run happily with 12
cores, 24 threads, and I do not see the scaling issues we observed between
16 and 20 cores in my tests years ago.

-- 
  Thor Lancelot Simon	                                     tls%panix.com@localhost

    "The dirtiest word in art is the C-word.  I can't even say 'craft'
     without feeling dirty."	-Chuck Close