On Thu, Feb 25, 2021 at 4:57 PM Cameron Smith <[email protected]
<mailto:[email protected]>> wrote:
Hello,
Bringing this thread back from the dead...
We made progress with creation of a distributed dmplex that matches our
source mesh partition and are in need of help writing values into a
vector created from the dmplex object.
As discussed previously, we have created a DMPlex instance using:
DMPlexCreateFromCellListPetsc(...)
DMGetPointSF(...)
PetscSFSetGraph(...)
which gives us a distribution of mesh vertices and elements in the DM
object that matches the element-based partition of our unstructured
mesh.
We then mark mesh vertices on the geometric model boundary using
DMSetLabelValue(...) and a map from our mesh vertices to dmplex points
(created during dmplex definition of vtx coordinates).
Following this, we create a section for vertices:
> DMPlexGetDepthStratum(dm, 0, &vStart, &vEnd);
> PetscSectionCreate(PetscObjectComm((PetscObject) dm), &s);
> PetscSectionSetNumFields(s, 1);
> PetscSectionSetFieldComponents(s, 0, 1);
> PetscSectionSetChart(s, vStart, vEnd);
> for(PetscInt v = vStart; v < vEnd; ++v) {
> PetscSectionSetDof(s, v, 1);
> PetscSectionSetFieldDof(s, v, 0, 1);
> }
> PetscSectionSetUp(s);
> DMSetLocalSection(dm, s);
> PetscSectionDestroy(&s);
> DMGetGlobalSection(dm,&s); //update the global section
We then try to write values into a local Vec for the on-process
vertices
(roots and leaves in sf terms) and hit an ordering problem.
Specifically, we make the following sequence of calls:
DMGetLocalVector(dm,&bloc);
VecGetArrayWrite(bloc, &bwrite);
//for loop to write values to bwrite
VecRestoreArrayWrite(bloc, &bwrite);
DMLocalToGlobal(dm,bloc,INSERT_VALUES,b);
DMRestoreLocalVector(dm,&bloc);
There is an easy way to get diagnostics here. For the local vector
DMGetLocalSection(dm, &s);
PetscSectionGetOffset(s, v, &off);
will give you the offset into the array you got from VecGetArrayWrite()
for that vertex. You can get this wrapped up using DMPlexPointLocalWrite()
which is what I tend to use for this type of stuff.
For the global vector
DMGetGlobalSection(dm, &gs);
PetscSectionGetOffset(gs, v, &off);
will give you the offset into the portion of the global array that is
stored in this process.
If you do not own the values for this vertex, the number is negative,
and it is actually
-(i+1) if the index i is the valid one on the owning process.
Visualizing Vec 'b' in paraview, and the
original mesh, tells us that the dmplex topology and geometry (the
vertex coordinates) are correct but that the order we write values is
wrong (not total garbage... but clearly shifted).
We do not make any guarantees that global orders match local orders.
However, by default
we number up global unknowns in rank order, leaving out the dofs that we
not owned.
Does this make sense?
Thanks,
Matt
Is there anything obviously wrong in our described approach? I suspect
the section creation is wrong and/or we don't understand the order of
entries in the array returned by VecGetArrayWrite.
Please let us know if other info is needed. We are happy to share the
relevant source code.
Thank-you,
Cameron
On 8/25/20 8:34 AM, Cameron Smith wrote:
> On 8/24/20 4:57 PM, Matthew Knepley wrote:
>> On Mon, Aug 24, 2020 at 4:27 PM Jed Brown <[email protected]
<mailto:[email protected]>
>> <mailto:[email protected] <mailto:[email protected]>>> wrote:
>>
>> Cameron Smith <[email protected] <mailto:[email protected]>
<mailto:[email protected] <mailto:[email protected]>>> writes:
>>
>> > We made some progress with star forest creation but still
have
>> work to do.
>> >
>> > We revisited DMPlexCreateFromCellListParallelPetsc(...)
and got it
>> > working by sequentially partitioning the vertex
coordinates across
>> > processes to satisfy the 'vertexCoords' argument.
Specifically,
>> rank 0
>> > has the coordinates for vertices with global id 0:N/P-1,
rank 1
>> has
>> > N/P:2*(N/P)-1, and so on (N is the total number of global
>> vertices and P
>> > is the number of processes).
>> >
>> > The consequences of the sequential partition of vertex
>> coordinates in
>> > subsequent solver operations is not clear. Does it make
process i
>> > responsible for computations and communications
associated with
>> global
>> > vertices i*(N/P):(i+1)*(N/P)-1 ? We assumed it does and
wanted
>> to confirm.
>>
>> Yeah, in the sense that the corners would be owned by the
rank you
>> place them on.
>>
>> But many methods, especially high-order, perform assembly via
>> non-overlapping partition of elements, in which case the
>> "computations" happen where the elements are (with any required
>> vertex data for the closure of those elements being sent to
the rank
>> handling the element).
>>
>> Note that a typical pattern would be to create a parallel DMPlex
>> with a naive distribution, then repartition/distribute it.
>>
>>
>> As Jed says, CreateParallel() just makes the most naive
partition of
>> vertices because we have no other information. Once
>> the mesh is made, you call DMPlexDistribute() again to reduce
the edge
>> cut.
>>
>> Thanks,
>>
>> Matt
>>
>
>
> Thank you.
>
> This is being used for PIC code with low order 2d elements whose
mesh is
> partitioned to minimize communications during particle
operations. This
> partition will not be ideal for the field solve using petsc so we're
> exploring alternatives that will require minimal data movement
between
> the two partitions. Towards that, we'll keep pursuing the SF
creation.
>
> -Cameron
>
--
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which
their experiments lead.
-- Norbert Wiener
https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>
