- simplify procAddressing read/write
- avoid accessing points in faMeshReconstructor.
Can rely on the patch meshPoints (labelList), which does not need
access to a pointField
- report number of points on decomposed mesh.
Can be useful additional information.
Additional statistics for finite area decomposition
- provide bundled reconstructAllFields for various reconstructors
- remove reconstructPar checks for very old face addressing
(from foam2.0 - ie, older than OpenFOAM itself)
- bundle all reading into fieldsDistributor tools,
where it can be reused by various utilities as required.
- combine decomposition fields as respective fieldsCache
which eliminates most of the clutter from decomposePar
and similfies reuse in the future.
STYLE: remove old wordHashSet selection (deprecated in 2018)
BUG: incorrect face flip handling for faMeshReconstructor
- a latent bug which is not yet triggered since the faMesh faces are
currently only definable on boundary faces (which never flip)
- for indirect lists we use element-wise output streaming and read
back as a regular list. This approach cannot however work with
non-blocking mode - the receive buffers will simply not be filled
before attempting to read from them.
For contiguous data, the lowest overhead solution is to locally
flatten the indirect list and use the regular gather routines
for non-blocking mode. For non-contiguous data, can continue to
use the element-wise output, but cannot use non-blocking for it.
STYLE: use non-blocking consistently as default for globalIndex gather(s)
- most of the front-facing code was already using non-blocking,
but there were a few low-level routines defaulting to scheduled
(but never relied upon in the code).
- relocate templating to factory method 'New'.
Adds provisions for more general re-use.
- expose processor topology in globalMesh as topology()
- wrap proc->patch lookup as processorTopology::procPatchLookup method
(failsafe). May consider using Map<label> for its storage in the
future.
- additional Pstream::broadcasts() method to serialize/deserialize
multiple items.
- revoke the broadcast specialisations for std::string and List(s) and
use a generic broadcasting template. In most cases, the previous
specialisations would have required two broadcasts:
(1) for the size
(2) for the contiguous content.
Now favour reduced communication over potential local (intermediate)
storage that would have only benefited a few select cases.
ENH: refine PstreamBuffers access methods
- replace 'bool hasRecvData(label)' with 'label recvDataCount(label)'
to recover the number of unconsumed receive bytes from specified
processor. Can use 'labelList recvDataCounts()' to recover the
number of unconsumed receive bytes from all processor.
- additional peekRecvData() method (for transcribing contiguous data)
ENH: globalIndex whichProcID - check for isLocal first
- reasonable to assume that local items are searched for more
frequently, so do preliminary check for isLocal before performing
a more costly binary search of globalIndex offsets
ENH: masterUncollatedFileOperation - bundled scatter of status
- MPI_Gatherv requires contiguous data, but a byte-wise transfer can
quickly exceed the 'int' limits used for MPI sizes/offsets. Thus
gather label/scalar components when possible to increase the
effective size limit.
For non-contiguous types (or large contiguous data) now also
reverts to manual handling
ENH: handle contiguous data in GAMGAgglomeration gather values
- delegate to globalIndex::gatherValues static method (new)
- eliminate redundant size_ accounting
- drop extra 'Container' template parameter and replace functionality
with more flexible pack/unpack methods.
There is also a pack() method that handles indirect lists of lists
that can be used, for example, to pack a patch slice of faces.
Drop the 'operator()' method in favour of unpack to expose and properly
document the conversion. Should revisit the corresponding code in
some places for optimization potential.
- align some method names with globalIndex:
totalSize(), maxSize() etc
- less communication than gatherList/scatterList
ENH: refine send granularity in Pstream::exchange
STYLE: ensure PstreamBuffers and defaultCommsType agree
- simpler loops for lduSchedule
- now largely encapsulated using PstreamBuffers methods,
which makes it simpler to centralize and maintain
- avoid building intermediate structures when sending data,
remove unused methods/data
TUT: parallel version of depthCharge2D
STYLE: minor update in ProcessorTopology
- PstreamBuffers nProcs() and allProcs() methods to recover the rank
information consistent with the communicator used for construction
- allowClearRecv() methods for more control over buffer reuse
For example,
pBufs.allowClearRecv(false);
forAll(particles, particlei)
{
pBufs.clear();
fill...
read via IPstream(..., pBufs);
}
This preserves the receive buffers memory allocation between calls.
- finishedNeighbourSends() method as compact wrapper for
finishedSends() when send/recv ranks are identically
(eg, neighbours)
- hasSendData()/hasRecvData() methods for PstreamBuffers.
Can be useful for some situations to skip reading entirely.
For example,
pBufs.finishedNeighbourSends(neighProcs);
if (!returnReduce(pBufs.hasRecvData(), orOp<bool>()))
{
// Nothing to do
continue;
}
...
On an individual basis:
for (const int proci : pBufs.allProcs())
{
if (pBufs.hasRecvData(proci))
{
...
}
}
Also conceivable to do the following instead (nonBlocking only):
if (!returnReduce(pBufs.hasSendData(), orOp<bool>()))
{
// Nothing to do
pBufs.clear();
continue;
}
pBufs.finishedNeighbourSends(neighProcs);
...
- reduces later surprises and simplifies effort for the caller
- more flexible globalIndex scatter with auto-sized return field.
- Avoid communication for scattering into zero-sized fields.
GIT: relocate globalIndex (is independent of mesh)
STYLE: include label/scalar Fwd in contiguous.H
STYLE: unneed commSchedule include in GeometricField
