Don't worry about the depth. Even if you have less levels of caches in
some parts of the machine, or no hyperthreading or whatever different,
we still build levels of identical objects. So you still get a
consistent depth for all cores, threads, sockets, and different levels
of caches. But obviously, you will see less objects when iterating
through levels within the "smaller" part of the machine.
If you want to run some tests, I can generate a fake heterogeneous
topology as a XML file that hwloc would load through the HWLOC_XMLFILE
Le 13/04/2011 19:21, Rayson Ho a écrit :
> Many thanks Brice!
> And I have a question for the list -- if the machine has processors
> that are not idential - ie. some "processors" have SMT (or HT)
> enabled, and some don't even have this support in the hardware, then
> what would the tree look like?? Or if the cache structure is different
> among the "processors", then would I still be able to use
> hwloc_get_depth_type() to get a consistent depth while I am iterating
> through the tree??
> So far, my code has only been tested on homogeneous machines.
> On Tue, Apr 12, 2011 at 4:21 PM, Brice Goglin <Brice.Goglin_at_[hidden]> wrote:
>> Le 12/04/2011 22:12, Rayson Ho a écrit :
>>> Can someone please add "Open Grid Scheduler" to the list of "...
>>> software already benefit from hwloc or are being ported to it" in the
>>> hwloc project homepage??
>>> Our homepage is:
>>> (Open Grid Scheduler is a fork of Sun Grid Engine, in total there are
>>> 3 forks. And Oracle is still developing its own closed source
>>> We will update the project status at:
>>> We still need to fix a bug or 2, and then we will ship hwloc support
>>> in SGE 6.2u5p3.
>>> hwloc-devel mailing list
> hwloc-devel mailing list