[MOAB-dev] iMeshP function argument

Tim Tautges tautges at mcs.anl.gov
Thu Jun 16 10:57:09 CDT 2011


[also cc'ing itaps-dev...]

On 06/16/2011 10:48 AM, Hong-Jun Kim wrote:
> I have one more question about the argument, command_code, as in the
> following function.
>
> void iMeshP_exchEntArrToPartsAll(
> iMesh_Instance instance,
> const iMeshP_PartitionHandle partition,
> const iBase_EntityHandle *entities,
> const int entities_size,
> const iMeshP_Part *target_part_ids,
> int command_code,
> int update_ghost,
> iMeshP_RequestHandle *request,
> int *err);
>
> The command_code is indicating whether to migrate input entities or only
> makes copies, that is, the ownership should be changed if it is set as 1.
>
> However, this option can make some questions to me.
>
> At first, if the entities are migrated to another processor, should the
> entities be deleted in sending processor?
>

I don't think so.

> In addition, if there are already shared information in entities or the
> exchanged entities are multi-shared, I can not decide which processor
> has the ownership.
> For example, if the entity is shared by p0, p1 and p2 and they are sent
> from p0 to p1 and p2, which processor should have the ownership?
>
> And, if the entity's ownership is already got from p0, when the entity
> is sent from p1 to only p2, the entity's ownership is changed to p2 in
> p1 and p2, and should the ownership in p0 also be changed to p2?
>

Good questions.  I think RPI is the only institution who's implemented this function so far, so maybe they can comment?

- tim

> I think there can be more problems to me when command_code is set to 1,
> and then can you give me some comments about it not to wander about the
> function?
>
> Thanks.
>
> Hong-Jun
>
>
> ------------------------------------------------------------------------
>
>  > From: "Tim Tautges" <tautges at mcs.anl.gov>
>  > To: moab-dev at mcs.anl.gov
>  > Sent: Wednesday, June 15, 2011 3:07:36 PM
>  > Subject: Re: [MOAB-dev] iMeshP function argument
>  > I think in this case the number of part ids is the same as the number
>  > of entities.
>  >
>  > - tim
>  >
>  > On 06/14/2011 02:20 PM, Hong-Jun Kim wrote:
>  > > Hi, all
>  > >
>  > > I am implementing the following function in iMeshP, but I just want
>  > > to
>  > > confirm "const iMeshP_Part *target_part_ids" argument is correct.
>  > >
>  > > I think it should be "const iMeshP_Part target_part_ids" for one
>  > > input
>  > > part, or for multiple input parts, there should be one more "const
>  > > int
>  > > part_id_size" argument.
>  > > Isn't it right?
>  > >
>  > > void iMeshP_exchEntArrToPartsAll(
>  > > iMesh_Instance instance,
>  > > const iMeshP_PartitionHandle partition,
>  > > const iBase_EntityHandle *entities,
>  > > const int entities_size,
>  > > const iMeshP_Part *target_part_ids,
>  > > int command_code,
>  > > int update_ghost,
>  > > iMeshP_RequestHandle *request,
>  > > int *err);
>  > >
>  > > -----------------------------
>  > > Hong-Jun Kim
>  > > Post-doc researcher
>  > > MCS, Argonne National Laboratory
>  > > 9700 S. Cass Ave. B240/R2147
>  > > Argonne, IL 60439
>  > > 630-252-4791
>  > > hongjun at mcs.anl.gov
>  > > -----------------------------
>  > >
>  >
>  > --
>  > ================================================================
>  > "You will keep in perfect peace him whose mind is
>  > steadfast, because he trusts in you." Isaiah 26:3
>  >
>  > Tim Tautges Argonne National Laboratory
>  > (tautges at mcs.anl.gov) (telecommuting from UW-Madison)
>  > phone: (608) 263-8485 1500 Engineering Dr.
>  > fax: (608) 263-4499 Madison, WI 53706

-- 
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (tautges at mcs.anl.gov)      (telecommuting from UW-Madison)
          phone: (608) 263-8485      1500 Engineering Dr.
            fax: (608) 263-4499      Madison, WI 53706



More information about the moab-dev mailing list