OpenFOAM-5.x/applications/utilities/surface/surfaceRedistributePar/surfaceRedistributePar.C
Henry Weller 7c301dbff4 Parallel IO: New collated file format
When an OpenFOAM simulation runs in parallel, the data for decomposed fields and
mesh(es) has historically been stored in multiple files within separate
directories for each processor.  Processor directories are named 'processorN',
where N is the processor number.

This commit introduces an alternative "collated" file format where the data for
each decomposed field (and mesh) is collated into a single file, which is
written and read on the master processor.  The files are stored in a single
directory named 'processors'.

The new format produces significantly fewer files - one per field, instead of N
per field.  For large parallel cases, this avoids the restriction on the number
of open files imposed by the operating system limits.

The file writing can be threaded allowing the simulation to continue running
while the data is being written to file.  NFS (Network File System) is not
needed when using the the collated format and additionally, there is an option
to run without NFS with the original uncollated approach, known as
"masterUncollated".

The controls for the file handling are in the OptimisationSwitches of
etc/controlDict:

OptimisationSwitches
{
    ...

    //- Parallel IO file handler
    //  uncollated (default), collated or masterUncollated
    fileHandler uncollated;

    //- collated: thread buffer size for queued file writes.
    //  If set to 0 or not sufficient for the file size threading is not used.
    //  Default: 2e9
    maxThreadFileBufferSize 2e9;

    //- masterUncollated: non-blocking buffer size.
    //  If the file exceeds this buffer size scheduled transfer is used.
    //  Default: 2e9
    maxMasterFileBufferSize 2e9;
}

When using the collated file handling, memory is allocated for the data in the
thread.  maxThreadFileBufferSize sets the maximum size of memory in bytes that
is allocated.  If the data exceeds this size, the write does not use threading.

When using the masterUncollated file handling, non-blocking MPI communication
requires a sufficiently large memory buffer on the master node.
maxMasterFileBufferSize sets the maximum size in bytes of the buffer.  If the
data exceeds this size, the system uses scheduled communication.

The installation defaults for the fileHandler choice, maxThreadFileBufferSize
and maxMasterFileBufferSize (set in etc/controlDict) can be over-ridden within
the case controlDict file, like other parameters.  Additionally the fileHandler
can be set by:
- the "-fileHandler" command line argument;
- a FOAM_FILEHANDLER environment variable.

A foamFormatConvert utility allows users to convert files between the collated
and uncollated formats, e.g.
    mpirun -np 2 foamFormatConvert -parallel -fileHandler uncollated

An example case demonstrating the file handling methods is provided in:
$FOAM_TUTORIALS/IO/fileHandling

The work was undertaken by Mattijs Janssens, in collaboration with Henry Weller.
2017-07-07 11:39:56 +01:00

295 lines
8.1 KiB
C

/*---------------------------------------------------------------------------*\
========= |
\\ / F ield | OpenFOAM: The Open Source CFD Toolbox
\\ / O peration |
\\ / A nd | Copyright (C) 2011-2017 OpenFOAM Foundation
\\/ M anipulation |
-------------------------------------------------------------------------------
License
This file is part of OpenFOAM.
OpenFOAM is free software: you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with OpenFOAM. If not, see <http://www.gnu.org/licenses/>.
Application
surfaceRedistributePar
Description
(Re)distribution of triSurface. Either takes an undecomposed surface
or an already decomposed surface and redistributes it so that each
processor has all triangles that overlap its mesh.
Note
- best decomposition option is hierarchGeomDecomp since
guarantees square decompositions.
- triangles might be present on multiple processors.
- merging uses geometric tolerance so take care with writing precision.
\*---------------------------------------------------------------------------*/
#include "argList.H"
#include "Time.H"
#include "polyMesh.H"
#include "distributedTriSurfaceMesh.H"
#include "mapDistribute.H"
#include "localIOdictionary.H"
using namespace Foam;
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
// Print on master all the per-processor surface stats.
void writeProcStats
(
const triSurface& s,
const List<List<treeBoundBox>>& meshBb
)
{
// Determine surface bounding boxes, faces, points
List<treeBoundBox> surfBb(Pstream::nProcs());
{
surfBb[Pstream::myProcNo()] = treeBoundBox(s.points());
Pstream::gatherList(surfBb);
Pstream::scatterList(surfBb);
}
labelList nPoints(Pstream::nProcs());
nPoints[Pstream::myProcNo()] = s.points().size();
Pstream::gatherList(nPoints);
Pstream::scatterList(nPoints);
labelList nFaces(Pstream::nProcs());
nFaces[Pstream::myProcNo()] = s.size();
Pstream::gatherList(nFaces);
Pstream::scatterList(nFaces);
forAll(surfBb, proci)
{
const List<treeBoundBox>& bbs = meshBb[proci];
Info<< "processor" << proci << nl
<< "\tMesh bounds : " << bbs[0] << nl;
for (label i = 1; i < bbs.size(); i++)
{
Info<< "\t " << bbs[i]<< nl;
}
Info<< "\tSurface bounding box : " << surfBb[proci] << nl
<< "\tTriangles : " << nFaces[proci] << nl
<< "\tVertices : " << nPoints[proci]
<< endl;
}
Info<< endl;
}
int main(int argc, char *argv[])
{
argList::addNote
(
"redistribute a triSurface"
);
argList::validArgs.append("surface file");
argList::validArgs.append("distribution type");
argList::addBoolOption
(
"keepNonMapped",
"preserve surface outside of mesh bounds"
);
#include "setRootCase.H"
#include "createTime.H"
runTime.functionObjects().off();
const fileName surfFileName = args[1];
const word distType = args[2];
Info<< "Reading surface from " << surfFileName << nl << nl
<< "Using distribution method "
<< distributedTriSurfaceMesh::distributionTypeNames_[distType]
<< " " << distType << nl << endl;
const bool keepNonMapped = args.options().found("keepNonMapped");
if (keepNonMapped)
{
Info<< "Preserving surface outside of mesh bounds." << nl << endl;
}
else
{
Info<< "Removing surface outside of mesh bounds." << nl << endl;
}
if (!Pstream::parRun())
{
FatalErrorInFunction
<< "Please run this program on the decomposed case."
<< " It will read surface " << surfFileName
<< " and decompose it such that it overlaps the mesh bounding box."
<< exit(FatalError);
}
#include "createPolyMesh.H"
Random rndGen(653213);
// Determine mesh bounding boxes:
List<List<treeBoundBox>> meshBb(Pstream::nProcs());
{
meshBb[Pstream::myProcNo()] = List<treeBoundBox>
(
1,
treeBoundBox
(
boundBox(mesh.points(), false)
).extend(rndGen, 1e-3)
);
Pstream::gatherList(meshBb);
Pstream::scatterList(meshBb);
}
IOobject io
(
surfFileName, // name
//runTime.findInstance("triSurface", surfFileName), // instance
runTime.constant(), // instance
"triSurface", // local
runTime, // registry
IOobject::MUST_READ,
IOobject::AUTO_WRITE
);
// Look for file (using searchableSurface rules)
const fileName actualPath(typeFilePath<searchableSurface>(io));
fileName localPath(actualPath);
localPath.replace(runTime.rootPath() + '/', "");
if (actualPath == io.objectPath())
{
Info<< "Loading local (decomposed) surface " << localPath << nl <<endl;
}
else
{
Info<< "Loading undecomposed surface " << localPath << nl << endl;
}
// Create dummy dictionary for bounding boxes if does not exist.
if (!isFile(actualPath / "Dict"))
{
dictionary dict;
dict.add("bounds", meshBb[Pstream::myProcNo()]);
dict.add("distributionType", distType);
dict.add("mergeDistance", SMALL);
localIOdictionary ioDict
(
IOobject
(
io.name() + "Dict",
io.instance(),
io.local(),
io.db(),
IOobject::NO_READ,
IOobject::NO_WRITE,
false
),
dict
);
Info<< "Writing dummy bounds dictionary to " << ioDict.name()
<< nl << endl;
// Force writing in ascii
ioDict.regIOobject::writeObject
(
IOstream::ASCII,
IOstream::currentVersion,
ioDict.time().writeCompression(),
true
);
}
// Load surface
distributedTriSurfaceMesh surfMesh(io);
Info<< "Loaded surface" << nl << endl;
// Generate a test field
{
const triSurface& s = static_cast<const triSurface&>(surfMesh);
autoPtr<triSurfaceVectorField> fcPtr
(
new triSurfaceVectorField
(
IOobject
(
"faceCentres", // name
surfMesh.searchableSurface::time().timeName(), // instance
surfMesh,
IOobject::NO_READ,
IOobject::AUTO_WRITE
),
surfMesh,
dimLength,
s.faceCentres()
)
);
// Steal pointer and store object on surfMesh
fcPtr.ptr()->store();
}
// Write per-processor stats
Info<< "Before redistribution:" << endl;
writeProcStats(surfMesh, meshBb);
// Do redistribution
Info<< "Redistributing surface" << nl << endl;
autoPtr<mapDistribute> faceMap;
autoPtr<mapDistribute> pointMap;
surfMesh.distribute
(
meshBb[Pstream::myProcNo()],
keepNonMapped,
faceMap,
pointMap
);
faceMap.clear();
pointMap.clear();
Info<< endl;
// Write per-processor stats
Info<< "After redistribution:" << endl;
writeProcStats(surfMesh, meshBb);
Info<< "Writing surface." << nl << endl;
surfMesh.objectRegistry::write();
Info<< "End\n" << endl;
return 0;
}
// ************************************************************************* //