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.
except turbulence and lagrangian which will also be updated shortly.
For example in the nonNewtonianIcoFoam offsetCylinder tutorial the viscosity
model coefficients may be specified in the corresponding "<type>Coeffs"
sub-dictionary:
transportModel CrossPowerLaw;
CrossPowerLawCoeffs
{
nu0 [0 2 -1 0 0 0 0] 0.01;
nuInf [0 2 -1 0 0 0 0] 10;
m [0 0 1 0 0 0 0] 0.4;
n [0 0 0 0 0 0 0] 3;
}
BirdCarreauCoeffs
{
nu0 [0 2 -1 0 0 0 0] 1e-06;
nuInf [0 2 -1 0 0 0 0] 1e-06;
k [0 0 1 0 0 0 0] 0;
n [0 0 0 0 0 0 0] 1;
}
which allows a quick change between models, or using the simpler
transportModel CrossPowerLaw;
nu0 [0 2 -1 0 0 0 0] 0.01;
nuInf [0 2 -1 0 0 0 0] 10;
m [0 0 1 0 0 0 0] 0.4;
n [0 0 0 0 0 0 0] 3;
if quick switching between models is not required.
To support this more convenient parameter specification the inconsistent
specification of seedSampleSet in the streamLine and wallBoundedStreamLine
functionObjects had to be corrected from
// Seeding method.
seedSampleSet uniform; //cloud; //triSurfaceMeshPointSet;
uniformCoeffs
{
type uniform;
axis x; //distance;
// Note: tracks slightly offset so as not to be on a face
start (-1.001 -0.05 0.0011);
end (-1.001 -0.05 1.0011);
nPoints 20;
}
to the simpler
// Seeding method.
seedSampleSet
{
type uniform;
axis x; //distance;
// Note: tracks slightly offset so as not to be on a face
start (-1.001 -0.05 0.0011);
end (-1.001 -0.05 1.0011);
nPoints 20;
}
which also support the "<type>Coeffs" form
// Seeding method.
seedSampleSet
{
type uniform;
uniformCoeffs
{
axis x; //distance;
// Note: tracks slightly offset so as not to be on a face
start (-1.001 -0.05 0.0011);
end (-1.001 -0.05 1.0011);
nPoints 20;
}
}
Avoids slight phase-fraction unboundedness at entertainment BCs and improved
robustness.
Additionally the phase-fractions in the multi-phase (rather than two-phase)
solvers are adjusted to avoid the slow growth of inconsistency ("drift") caused
by solving for all of the phase-fractions rather than deriving one from the
others.
Previously the inlet flow of phase 1 (the phase solved for) is corrected
to match the inlet specification for that phase. However, if the second
phase is also constrained at inlets the inlet flux must also be
corrected to match the inlet specification.
When the GeometricBoundaryField template class was originally written it
was a separate class in the Foam namespace rather than a sub-class of
GeometricField as it is now. Without loss of clarity and simplifying
code which access the boundary field of GeometricFields it is better
that GeometricBoundaryField be renamed Boundary for consistency with the
new naming convention for the type of the dimensioned internal field:
Internal, see commit a25a449c9e
This is a very simple text substitution change which can be applied to
any code which compiles with the OpenFOAM-dev libraries.
Given that the type of the dimensioned internal field is encapsulated in
the GeometricField class the name need not include "Field"; the type
name is "Internal" so
volScalarField::DimensionedInternalField -> volScalarField::Internal
In addition to the ".dimensionedInternalField()" access function the
simpler "()" de-reference operator is also provided to greatly simplify
FV equation source term expressions which need not evaluate boundary
conditions. To demonstrate this kEpsilon.C has been updated to use
dimensioned internal field expressions in the k and epsilon equation
source terms.
Resolves bug-report http://www.openfoam.org/mantisbt/view.php?id=1938
Because C++ does not support overloading based on the return-type there
is a problem defining both const and non-const member functions which
are resolved based on the const-ness of the object for which they are
called rather than the intent of the programmer declared via the
const-ness of the returned type. The issue for the "boundaryField()"
member function is that the non-const version increments the
event-counter and checks the state of the stored old-time fields in case
the returned value is altered whereas the const version has no
side-effects and simply returns the reference. If the the non-const
function is called within the patch-loop the event-counter may overflow.
To resolve this it in necessary to avoid calling the non-const form of
"boundaryField()" if the results is not altered and cache the reference
outside the patch-loop when mutation of the patch fields is needed.
The most straight forward way of resolving this problem is to name the
const and non-const forms of the member functions differently e.g. the
non-const form could be named:
mutableBoundaryField()
mutBoundaryField()
nonConstBoundaryField()
boundaryFieldRef()
Given that in C++ a reference is non-const unless specified as const:
"T&" vs "const T&" the logical convention would be
boundaryFieldRef()
boundaryFieldConstRef()
and given that the const form which is more commonly used is it could
simply be named "boundaryField()" then the logical convention is
GeometricBoundaryField& boundaryFieldRef();
inline const GeometricBoundaryField& boundaryField() const;
This is also consistent with the new "tmp" class for which non-const
access to the stored object is obtained using the ".ref()" member function.
This new convention for non-const access to the components of
GeometricField will be applied to "dimensionedInternalField()" and "internalField()" in the
future, i.e. "dimensionedInternalFieldRef()" and "internalFieldRef()".
e.g. (fvc::interpolate(HbyA) & mesh.Sf()) -> fvc::flux(HbyA)
This removes the need to create an intermediate face-vector field when
computing fluxes which is more efficient, reduces the peak storage and
improved cache coherency in addition to providing a simpler and cleaner
API.
The deprecated non-const tmp functionality is now on the compiler switch
NON_CONST_TMP which can be enabled by adding -DNON_CONST_TMP to EXE_INC
in the Make/options file. However, it is recommended to upgrade all
code to the new safer tmp by using the '.ref()' member function rather
than the non-const '()' dereference operator when non-const access to
the temporary object is required.
Please report any problems on Mantis.
Henry G. Weller
CFD Direct.
fvOptions are transferred to the database on construction using
fv::options::New which returns a reference. The same function can be
use for construction and lookup so that fvOptions are now entirely
demand-driven.
The abstract base-classes for fvOptions now reside in the finiteVolume
library simplifying compilation and linkage. The concrete
implementations of fvOptions are still in the single monolithic
fvOptions library but in the future this will be separated into smaller
libraries based on application area which may be linked at run-time in
the same manner as functionObjects.
This formulation provides C-grid like pressure-flux staggering on an
unstructured mesh which is hugely beneficial for Euler-Euler multiphase
equations as it allows for all forces to be treated in a consistent
manner on the cell-faces which provides better balance, stability and
accuracy. However, to achieve face-force consistency the momentum
transport terms must be interpolated to the faces reducing accuracy of
this part of the system but this is offset by the increase in accuracy
of the force-balance.
Currently it is not clear if this face-based momentum equation
formulation is preferable for all Euler-Euler simulations so I have
included it on a switch to allow evaluation and comparison with the
previous cell-based formulation. To try the new algorithm simply switch
it on, e.g.:
PIMPLE
{
nOuterCorrectors 3;
nCorrectors 1;
nNonOrthogonalCorrectors 0;
faceMomentum yes;
}
It is proving particularly good for bubbly flows, eliminating the
staggering patterns often seen in the air velocity field with the
previous algorithm, removing other spurious numerical artifacts in the
velocity fields and improving stability and allowing larger time-steps
For particle-gas flows the advantage is noticeable but not nearly as
pronounced as in the bubbly flow cases.
Please test the new algorithm on your cases and provide feedback.
Henry G. Weller
CFD Direct
