dolfin team mailing list archive

["Garth N. Wells" <gnw20@xxxxxxxxx>]

On Mon, 2010-08-09 at 17:39 +0200, Anders Logg wrote:
> On Mon, Aug 09, 2010 at 04:04:24PM +0100, Garth N. Wells wrote:
> > On Mon, 2010-08-09 at 16:54 +0200, Anders Logg wrote:
> > > On Mon, Aug 09, 2010 at 03:47:17PM +0100, Garth N. Wells wrote:
> > > > On Mon, 2010-08-09 at 16:41 +0200, Anders Logg wrote:
> > > > > On Mon, Aug 09, 2010 at 03:30:03PM +0100, Garth N. Wells wrote:
> > > > > > On Mon, 2010-08-09 at 16:24 +0200, Anders Logg wrote:
> > > > > > > On Mon, Aug 09, 2010 at 03:21:59PM +0100, Garth N. Wells wrote:
> > > > > > > > On Mon, 2010-08-09 at 16:14 +0200, Anders Logg wrote:
> > > > > > > > > On Mon, Aug 09, 2010 at 02:34:45PM +0100, Garth N. Wells wrote:
> > > > > > > > > > On Mon, 2010-08-09 at 15:20 +0200, Anders Logg wrote:
> > > > > > > > > > > On Mon, Aug 09, 2010 at 02:12:49PM +0100, Garth N. Wells wrote:
> > > > > > > > > > > > On Mon, 2010-08-09 at 15:05 +0200, Anders Logg wrote:
> > > > > > > > > > > > > On Mon, Aug 09, 2010 at 01:54:04PM +0100, Garth N. Wells wrote:
> > > > > > > > > > > > > > On Mon, 2010-08-09 at 14:46 +0200, Anders Logg wrote:
> > > > > > > > > > > > > > > On Mon, Aug 09, 2010 at 01:09:47PM +0100, Garth N. Wells wrote:
> > > > > > > > > > > > > > > > On Mon, 2010-08-09 at 13:53 +0200, Anders Logg wrote:
> > > > > > > > > > > > > > > > > On Mon, Aug 09, 2010 at 12:47:10PM +0100, Garth N. Wells wrote:
> > > > > > > > > > > > > > > > > > On Mon, 2010-08-09 at 13:37 +0200, Anders Logg wrote:
> > > > > > > > > > > > > > > > > > > On Sat, Aug 07, 2010 at 01:24:44PM +0100, Garth N. Wells wrote:
> > > > > > > > > > > > > > > > > > > > On Fri, 2010-08-06 at 19:55 +0100, Garth N. Wells wrote:
> > > > > > > > > > > > > > > > > > > > > On Fri, 2010-08-06 at 20:53 +0200, Anders Logg wrote:
> > > > > > > > > > > > > > > > > > > > > > On Fri, Aug 06, 2010 at 07:51:18PM +0100, Garth N. Wells wrote:
> > > > > > > > > > > > > > > > > > > > > > > On Fri, 2010-08-06 at 20:36 +0200, Anders Logg wrote:
> > > > > > > > > > > > > > > > > > > > > > > > On Fri, Aug 06, 2010 at 04:55:44PM +0100, Garth N. Wells wrote:
> > > > > > > > > > > > > > > > > > > > > > > > > On Fri, 2010-08-06 at 08:42 -0700, Johan Hake wrote:
> > > > > > > > > > > > > > > > > > > > > > > > > > On Friday August 6 2010 08:16:26 you wrote:
> > > > > > > > > > > > > > > > > > > > > > > > > > > ------------------------------------------------------------
> > > > > > > > > > > > > > > > > > > > > > > > > > > revno: 4896
> > > > > > > > > > > > > > > > > > > > > > > > > > > committer: Garth N. Wells <gnw20@xxxxxxxxx>
> > > > > > > > > > > > > > > > > > > > > > > > > > > branch nick: dolfin-all
> > > > > > > > > > > > > > > > > > > > > > > > > > > timestamp: Fri 2010-08-06 16:13:29 +0100
> > > > > > > > > > > > > > > > > > > > > > > > > > > message:
> > > > > > > > > > > > > > > > > > > > > > > > > > >   Add simple Stokes solver for parallel testing.
> > > > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > > > >   Other Stokes demos don't run in parallel because MeshFunction io is not
> > > > > > > > > > > > > > > > > > > > > > > > > > >   supported in parallel.
> > > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > > > Does anyone have an overview of what is needed for this to be fixed. I
> > > > > > > > > > > > > > > > > > > > > > > > > > couldn't find a blueprint on it.
> > > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > > Here it is:
> > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > >     https://blueprints.launchpad.net/dolfin/+spec/parallel-io
> > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > > > I am interested in getting this fixed :)
> > > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > > Me too! We need to look at all the io since much of it is broken in
> > > > > > > > > > > > > > > > > > > > > > > > > parallel.
> > > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > > We need to settle on how to handle XML data. I favour (and I know Niclas
> > > > > > > > > > > > > > > > > > > > > > > > > Janson does too) the VTK approach in which we have a 'master file' that
> > > > > > > > > > > > > > > > > > > > > > > > > points to other XML files which contain portions of the vector/mesh,
> > > > > > > > > > > > > > > > > > > > > > > > > etc. Process zero can read the 'master file' and then instruct the other
> > > > > > > > > > > > > > > > > > > > > > > > > processes on which file(s) they should read in.
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > This only works if the data is already partitioned. Most of our demos
> > > > > > > > > > > > > > > > > > > > > > > > assume that we have the mesh in one single file which is then
> > > > > > > > > > > > > > > > > > > > > > > > partitioned on the fly.
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > The approach does work for data which is not partitioned. Just like with
> > > > > > > > > > > > > > > > > > > > > > > VTK, one can read the 'master file' or the individual files.
> > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > The initial plan was to support two different ways of reading data in parallel:
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > 1. One file and automatic partitioning
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > DOLFIN gets one file "mesh.xml", each process reads one part of it (just
> > > > > > > > > > > > > > > > > > > > > > > > skipping other parts of the file), then the mesh is partitioned and
> > > > > > > > > > > > > > > > > > > > > > > > redistributed.
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > 2. Several files and no partitioning
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > DOLFIN get multiple files and each process reads one part. In this
> > > > > > > > > > > > > > > > > > > > > > > > case, the mesh and all associated data is already partitioned. This
> > > > > > > > > > > > > > > > > > > > > > > > should be very easy to fix since everything that is needed is already
> > > > > > > > > > > > > > > > > > > > > > > > in place; we just need to fix the logic. In particular, the data
> > > > > > > > > > > > > > > > > > > > > > > > section of each local mesh contains all auxilliary parallel data.
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > This can be handled in two different ways. Either a user specifies the
> > > > > > > > > > > > > > > > > > > > > > > > name of the file as "mesh*.xml", in which case DOLFIN appends say
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > >   "_%d" % MPI::process_number()
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > on each local process.
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > > The other way is to have a master file which lists all the other
> > > > > > > > > > > > > > > > > > > > > > > > files. In this case, I don't see a need for process 0 to take any kind
> > > > > > > > > > > > > > > > > > > > > > > > of responsibility for communicating file names. It would work fine for
> > > > > > > > > > > > > > > > > > > > > > > > each process to read the master file and then check which file it
> > > > > > > > > > > > > > > > > > > > > > > > should use. Each process could also check that the total number of
> > > > > > > > > > > > > > > > > > > > > > > > processes matches the number of partitions in the file. We could let
> > > > > > > > > > > > > > > > > > > > > > > > process 0 handle the parsing of the master file and then communicate
> > > > > > > > > > > > > > > > > > > > > > > > the file names but maybe that is an extra complication.
> > > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > > This fails when the number of files differs from the number of
> > > > > > > > > > > > > > > > > > > > > > > processes. It's very important to support m files on n processes. We've
> > > > > > > > > > > > > > > > > > > > > > > discussed this at length before.
> > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > > I don't remember. Can you remind me of what the reasons are?
> > > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > > I perform a simulation using m processes, and write the result to m
> > > > > > > > > > > > > > > > > > > > > files. Later I want to use the result later in another computation using
> > > > > > > > > > > > > > > > > > > > > n processors.
> > > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > I've looked a little into parallel io, and looked at what Trilinos and
> > > > > > > > > > > > > > > > > > > > PETSc do. Both support HDF5, and HDF5 has been developed to work
> > > > > > > > > > > > > > > > > > > > parallel. HDF5 does not advocate the one-file per processes (too awkward
> > > > > > > > > > > > > > > > > > > > and complicated they say), but advocates a one file approach. It has
> > > > > > > > > > > > > > > > > > > > tools that allow different processes to write to different parts of the
> > > > > > > > > > > > > > > > > > > > same file in parallel.
> > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > >From reading this, what I propose (for now) is:
> > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > 1. We only ever write one XML file for a given object. This file can be
> > > > > > > > > > > > > > > > > > > > read by different processes, with each reading in only a chunk.
> > > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > > 2. We should add an XML format for partitioning data (Trilinos calls
> > > > > > > > > > > > > > > > > > > > this a 'map'). If a map file is present, it is used to define the
> > > > > > > > > > > > > > > > > > > > partitions. It may make sense to have a map file for each process (but
> > > > > > > > > > > > > > > > > > > > no need for a 'master file').
> > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > I suggest something slightly different. I'm ok with the one file
> > > > > > > > > > > > > > > > > > > approach, but it would be good to store that data in a partitioned
> > > > > > > > > > > > > > > > > > > way. Our current model for parallel computing is that each process has
> > > > > > > > > > > > > > > > > > > a Mesh and each process has the partitioning data it needs stored in
> > > > > > > > > > > > > > > > > > > the data section of the Mesh. So each process has just a regular mesh
> > > > > > > > > > > > > > > > > > > with some auxilliary data attached to it. That makes it easy to read
> > > > > > > > > > > > > > > > > > > and write using already existing code. (No need for a special parallel
> > > > > > > > > > > > > > > > > > > format.)
> > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > But we could easily throw all that data into one big file, something
> > > > > > > > > > > > > > > > > > > like this:
> > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > > <distributed_mesh num_parts="16">
> > > > > > > > > > > > > > > > > > >   <mesh ...>
> > > > > > > > > > > > > > > > > > >     ...
> > > > > > > > > > > > > > > > > > >   </mesh>
> > > > > > > > > > > > > > > > > > >   <mesh ...>
> > > > > > > > > > > > > > > > > > >     ...
> > > > > > > > > > > > > > > > > > >   </mesh>
> > > > > > > > > > > > > > > > > > >   ...
> > > > > > > > > > > > > > > > > > > </distributed_mesh>
> > > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > > I would like to separate mesh and partitioning data. A partitioning of a
> > > > > > > > > > > > > > > > > > given mesh is not unique to a mesh, so it should be separated. A
> > > > > > > > > > > > > > > > > > partition could still go in the same XML file though.
> > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > It is separate from the mesh, but it is stored locally as part of the
> > > > > > > > > > > > > > > > > Mesh in MeshData. I think this has proved to be a very good way
> > > > > > > > > > > > > > > > > (efficient and simple) to store the data.
> > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > This breaks the concept of using a given Mesh on a different number of
> > > > > > > > > > > > > > > > processes. For a given mesh, I may want to store and use different
> > > > > > > > > > > > > > > > partitions.
> > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > The same applied to Vectors.
> > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > > Or do you suggest that we store it differently on file than we do as
> > > > > > > > > > > > > > > > > part of the DOLFIN data structures?
> > > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > > I suggest that we decouple the Mesh and the partition data for file
> > > > > > > > > > > > > > > > output. It comes back to permitting a variable number of processes.
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > I thought that the storage output from m processes would be the mesh
> > > > > > > > > > > > > > > partitioned into m pieces with auxilliary partitioning
> > > > > > > > > > > > > > > information. This would then be read by n processes and repartitioned.
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > > > If you don't want the partitioning into a specific number of
> > > > > > > > > > > > > > > partitions (like m partitions above), then what extra partitioning
> > > > > > > > > > > > > > > data would be required? It would just be a regular mesh file.
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > It would map entities to processes. It could also be used for Vectors.
> > > > > > > > > > > > >
> > > > > > > > > > > > > Then how is that not coupled to a specific number of processes?
> > > > > > > > > > > > >
> > > > > > > > > > > > > And why would it be better than what we have today, which stores all
> > > > > > > > > > > > > the parallel data we need (not just that mapping but everything else
> > > > > > > > > > > > > that can and must be generated from it) as part of each mesh?
> > > > > > > > > > > > >
> > > > > > > > > > > > > > > Partitioning data is always coupled to a specific number of processes
> > > > > > > > > > > > > > > as far as I can imagine.
> > > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > Naturally, and so it should be. But the mesh output doesn't have to be,
> > > > > > > > > > > > > > which why I suggest having a Mesh and partitioning data, e.g.
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >   // Let DOLFIN partition and distribute the mesh
> > > > > > > > > > > > > >   Mesh meshA("mesh.xml")
> > > > > > > > > > > > >
> > > > > > > > > > > > > Is this just one regular mesh file?
> > > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > Yes, and ParMETIS would partition the mesh.
> > > > > > > > > > > >
> > > > > > > > > > > > > >   // Partition the mesh according to my_partition.xml mesh. Throw an
> > > > > > > > > > > > > >   // error if there is a mis-match between my_partition and the number
> > > > > > > > > > > > > >   // of processes
> > > > > > > > > > > > > >   Mesh meshB("mesh.xml", "my_partition.xml")
> > > > > > > > > > > > >
> > > > > > > > > > > > > I don't understand this use-case. The first line above would already
> > > > > > > > > > > > > partition the mesh (using ParMETIS), and then you want to repartition
> > > > > > > > > > > > > it. Is it the case that my_partition.xml contains a better
> > > > > > > > > > > > > partitioning than what ParMETIS can compute?
> > > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > In the case "meshB("mesh.xml", "my_partition.xml")", ParMETIS would
> > > > > > > > > > > > never be called. The mesh would be distributed according to
> > > > > > > > > > > > "my_partition.xml".
> > > > > > > > > > >
> > > > > > > > > > > I missed that you created two different meshes. ok, that sounds like a
> > > > > > > > > > > good thing to have.
> > > > > > > > > > >
> > > > > > > > > > > But how would the mesh.xml be stored on file? Just one big file?
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > Yes, just as we have now.
> > > > > > > > >
> > > > > > > > > ok.
> > > > > > > > >
> > > > > > > > > > > And how is the above example related to the storing of *partitioned*
> > > > > > > > > > > meshes?
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > >From reading around, I don't think that we should store partitioned
> > > > > > > > > > meshes in the sense that each partition is a separate file. We should
> > > > > > > > > > store an entire mesh in one file, and store the partitioning data if
> > > > > > > > > > desired.
> > > > > > > > >
> > > > > > > > > I don't understand why not. It's a good feature to have to be able to
> > > > > > > > > restart a simulation (on the same number of processors) and completely
> > > > > > > > > avoid the partitioning step. Moreover, everything we need is already
> > > > > > > > > implemented since we can read/write meshes including mesh data.
> > > > > > > > >
> > > > > > > >
> > > > > > > > It is a good feature. To do it, save the mesh and the partition, and
> > > > > > > > then restart using
> > > > > > > >
> > > > > > > >   Mesh mesh("mesh.xml", "partition.xml");
> > > > > > > >
> > > > > > > > This will not involve any partitioning.
> > > > > > >
> > > > > > > It's true that it wouldn't involve any partitioning (as in calling
> > > > > > > ParMETIS) but it would involve redistributing the mesh between all
> > > > > > > processes and computing all the auxilliary data that is needed, like
> > > > > > > the global vertex numbering and mappings to neighboring processes.
> > > > > >
> > > > > > No, because the process would only read the bits which it needs. The
> > > > > > partition data would contains the other data.
> > > > >
> > > > > So it would not just be a simple list of markers for which cells
> > > > > belong where? It would be all that other stuff that we stored in
> > > > > MeshData today?
> > > > >
> > > > > > > By storing one file per process (which is almost already supported),
> > > > > > > that would also be avoided. Each process would read a regular mesh
> > > > > > > file and be ready to go. No communication and no computation involved.
> > > > > > >
> > > > > > > > > In summary, if I understand correctly that you want things stored in
> > > > > > > > > one file, we would have the following:
> > > > > > > > >
> > > > > > > > > 1. Reading from one file and automatic partitioning/redistribution
> > > > > > > > >
> > > > > > > > > This is already supported today, and it already supports the use case
> > > > > > > > > of first running on m processors and then n processors since in both
> > > > > > > > > cases the mesh will be read from one single file and partitioned.
> > > > > > > > >
> > > > > > > >
> > > > > > > > Yes.
> > > > > > > >
> > > > > > > > > 2. Reading from multiple files where the mesh has already been
> > > > > > > > > partitioned and m = n. This is almost supported today. We just need to
> > > > > > > > > decide on the logic for reading from multiple files.
> > > > > > > > >
> > > > > > > >
> > > > > > > > I advocate not supporting multiple files. Keeping to one file-file model
> > > > > > > > makes things simple.
> > > > > > >
> > > > > > > Multiple files would not be much of a complication since it is almost
> > > > > > > already in place. Each process just reads its corresponding file
> > > > > > > without any communication.
> > > > > > >
> > > > > >
> > > > > > Having dug around in the XML files over the past few days, I wouldn't
> > > > > > agree with this. It may seem outwardly simple, but it means files having
> > > > > > different states and it gets messy.
> > > > >
> > > > > What do you mean by different states? Each process just reads a plain
> > > > > mesh XML file which is no different from a regular mesh file except
> > > > > that the mesh data would contain a few more things.
> > > > >
> > > >
> > > > In one case. But we also support reading/writing from/to one file or
> > > > multiple files. We need to decide who should open a file (is object
> > > > local or distributed), should DOLFIN or the user add a process suffix,
> > > > how should the object be created (which depends on how it's read in).
> > > > It's simpler (for the user and the developer) if there is just one way
> > > > to do something.
> > >
> > > How about the following:
> > >
> > > 1. Input/output is always just one file
> > >
> > > 2. Always assume parallel when MPI::num_processes() > 1
> > >
> > > 3. When reading a regular mesh file, it is always partitioned
> > >
> >
> > This round back-to-front. The majority of meshes read in won't already
> > be partitioned.
> >
> > > 4. The partitioning is either handled by ParMETIS, or alternatively,
> > > according to a file or MeshFunction<uint> specified by the user. In
> > > particular, the file containing the partition can just be a regular
> > > MeshFunction<uint> stored in XML.
> > >
> > > 5. When specifying the partition as above, the only difference from
> > > now is that the call to ParMETIS is bypassed. The rest remains the
> > > same.
> > >
> > > 6. When writing a mesh in parallel, we write something like this:
> > >
> > >   <distributed_mesh>
> > >     <mesh>
> > >
> > >     </mesh>
> > >     <mesh>
> > >
> > >     </mesh>
> > >     ...
> > >   </distributed_mesh>
> > >
> > > 7. When reading in parallel, DOLFIN will check if it gets a regular
> > > mesh file (in which case it partitions it like in 3) or if it gets a
> > > distributed mesh. In that case, it will check that the number of
> > > processors matches (m = n) and then just continue to let each process
> > > read the data it needs by just skipping to the appropriate location in
> > > the file and reading the portion it needs.
> > >
> > > That way, it would be very clear what happens. Clear to the user (by
> > > just inspecting the XML file, is it a mesh or a distributed mesh) and
> > > to DOLFIN.
> > >
> >
> > It doesn't seem that I'm making my point very clearly ;).
> 
> Perhaps not... ;-) Keep trying. At some point, I might see the light.
> 
> > A mesh is a mesh. When I open the mesh file I want to be able to just
> > read it, no matter how many processes I'm using or how it was previously
> > partitioned. This is why I object to partitioning data being stored in
> > the mesh. We shouldn't have <distributed_mesh>.
> 
> So you advocate just the same old mesh format we have today? (And
> possibly some additional HDF5 format later if/when we need it.)
> 

Yes.

> > If we have a MeshFuction which defines a partition, with this the mesh
> > XML reader knows which parts of the mesh file it is responsible for.
> 
> Yes, and I agree that's a good feature. My point is just this, if you
> want to rerun a simulation on m = n processors, you have three options:
> 
> 1. Just read from mesh.xml and let ParMETIS recompute the partitioning
> and redistribute the mesh.
> 

Yes.

> 2. Read from mesh.xml and supply partition.xml (storing a
> MeshFunction). DOLFIN can then use this to accomplish two things:
> 
>   (a) Each process reads exactly the data it needs (no redistribution necessary)
>   (b) No need to call ParMETIS
> 

Yes.

> We would still need to recompute all auxilliary parallel data (like
> the mapping to neighboring processes).
> 

This shouldn't be hard since the mesh.xml will contain the global
numbers.

> 3. Read from either one file containing <distribute_mesh> or from
> multiple files with plain meshes. This would by-pass all the above
> steps and be the most efficient step.
> 
> The points I want to make is that case (2) does not handle anything
> that case (3) does not handle; case (3) is the most efficient; and
> case (3) is (almost) already implemented so there are no complications
> involved in handling it.
> 
> My suggestion for handling case (3) is (as before) the following:
> 
>   Mesh mesh("mesh*.xml")
> 
> The * would indicate that we should read multiple files.
> 
> I'll be offline for a few hours so we can continue this lengthy
> discussion later tonight... :-)
> 

I come back to the point that a one-file approach is simpler. It's not
just me saying this, it's what I've gleaned from both reading around on
parallel io and from implementing things myself ;).

Garth

> --
> Anders