itaps-parallel Part handles

[Onkar Sahni]

>    + Part handles are globally unique (so that remote part handles make
> sense and Zoltan doesn't go insane);
> - remote part handles won't exist on a processor unless they were
> communicated there somehow; in that case the implementation will be able
> to store the proc # with (not in) the handle if it so chooses
> - didn't I hear you say, Karen, that Zoltan computes its own 0..n-1
> numbering of parts anyway?  Or, will you eliminate that if we have
> unique part handles?
> - are we assuming an application will allocate all the part handles
> before calling Zoltan for the partitioning case?  Does it actually have
> to migrate entities to the processor owning a part in order to assign
> those entities to that part?  Or, can a given processor only assign
> local entities to local parts?  That seems like a pretty severe
> limitation (and would only allow you to partition for increasing #'s of
> processors besides)

  Remote part-handles, for neighboring parts, would exist for a given part
(and they are likely communicated at the time when parts became
neighbors, and if communicated will be done by implementation). I
specifically asked (in a telecon) if applications would be allowed to
communicate part-handles around, and answer was no. And my concern is
how would Zoltan get access to part-handle (say PH10) on a local
part/process (say P1, and consider a part per process for now) which is
neither local nor neighbor (i.e., P10 is not neighbor to P1 and how can
one access PH10 on P1).

- Onkar