Le 17/02/2011 23:52, Jeff Squyres a écrit :
> On Feb 17, 2011, at 5:37 PM, Brice Goglin wrote:
>> I am waiting for this troll to settle down before pushing some related
>> changes (add a (cpubind?) support flag and a hwloc-bind option).
>> I agree that "last" is better than "current". And I agree that
>> "lastcpuexec" isn't perfect, but I didn't see anything better so far :)
> (you can't know your binding and location at the same time -- nyuk nyuk)
This one has one advantage: if we add another function to return the
memory location, the name pattern will match (get_last_memory_location).
Could be misleading.
These are ok
Not "my" (we can do it for other processes too).