openfoam/applications/utilities/preProcessing/changeDictionary/changeDictionary.C
Andrew Heather d8d6030ab6 INT: Integration of Mattijs' collocated parallel IO additions
Original commit message:
------------------------

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

721 lines
21 KiB
C

/*---------------------------------------------------------------------------*\
========= |
\\ / F ield | OpenFOAM: The Open Source CFD Toolbox
\\ / O peration |
\\ / A nd | Copyright (C) 2011-2016 OpenFOAM Foundation
\\/ M anipulation | Copyright (C) 2016 OpenCFD Ltd.
-------------------------------------------------------------------------------
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
changeDictionary
Group
grpPreProcessingUtilities
Description
Utility to change dictionary entries, e.g. can be used to change the patch
type in the field and polyMesh/boundary files.
Reads dictionaries (fields) and entries to change from a dictionary.
E.g. to make the \em movingWall a \em fixedValue for \em p but all other
\em Walls a zeroGradient boundary condition, the
\c system/changeDictionaryDict would contain the following:
\verbatim
p // field to change
{
boundaryField
{
".*Wall" // entry to change
{
type zeroGradient;
}
movingWall // entry to change
{
type fixedValue;
value uniform 123.45;
}
}
}
\endverbatim
Replacement entries starting with '~' will remove the entry.
Usage
\b changeDictionary [OPTION]
Options:
- \par -subDict
Specify the subDict name of the replacements dictionary.
- \par -literalRE
Do not interpret regular expressions or patchGroups; treat them as any
other keyword.
- \par -enableFunctionEntries
Enable function entries (default: disabled)
- \par -disablePatchGroups
Disable the default checking for keys being patchGroups
\*---------------------------------------------------------------------------*/
#include "argList.H"
#include "IOobjectList.H"
#include "IOPtrList.H"
#include "volFields.H"
#include "stringListOps.H"
#include "timeSelector.H"
using namespace Foam;
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
namespace Foam
{
defineTemplateTypeNameAndDebug(IOPtrList<entry>, 0);
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
// Extract groupPatch info from boundary file info
HashTable<wordList, word> extractPatchGroups(const dictionary& boundaryDict)
{
HashTable<wordList, word> groupToPatch;
forAllConstIter(dictionary, boundaryDict, iter)
{
const word& patchName = iter().keyword();
const dictionary& patchDict = iter().dict();
wordList groups;
if (patchDict.readIfPresent("inGroups", groups))
{
forAll(groups, i)
{
HashTable<wordList, word>::iterator fndGroup = groupToPatch.find
(
groups[i]
);
if (fndGroup == groupToPatch.end())
{
groupToPatch.insert(groups[i], wordList(1, patchName));
}
else
{
fndGroup().append(patchName);
}
}
}
}
return groupToPatch;
}
bool merge
(
const bool addNonExisting,
dictionary&,
const dictionary&,
const bool,
const HashTable<wordList, word>&
);
// Add thisEntry to dictionary thisDict.
bool addEntry
(
dictionary& thisDict,
entry& thisEntry,
const entry& mergeEntry,
const bool literalRE,
const HashTable<wordList, word>& shortcuts
)
{
bool changed = false;
// Recursively merge sub-dictionaries
// TODO: merge without copying
if (thisEntry.isDict() && mergeEntry.isDict())
{
if
(
merge
(
true,
const_cast<dictionary&>(thisEntry.dict()),
mergeEntry.dict(),
literalRE,
shortcuts
)
)
{
changed = true;
}
}
else
{
// Should use in-place modification instead of adding
thisDict.add(mergeEntry.clone(thisDict).ptr(), true);
changed = true;
}
return changed;
}
// List of indices into thisKeys
labelList findMatches
(
const HashTable<wordList, word>& shortcuts,
const wordList& shortcutNames,
const wordList& thisKeys,
const keyType& key
)
{
labelList matches;
if (key.isPattern())
{
// Wildcard match
matches = findStrings(key, thisKeys);
}
else if (shortcuts.size())
{
// See if patchGroups expand to valid thisKeys
labelList indices = findStrings(key, shortcutNames);
forAll(indices, i)
{
const word& name = shortcutNames[indices[i]];
const wordList& keys = shortcuts[name];
forAll(keys, j)
{
label index = findIndex(thisKeys, keys[j]);
if (index != -1)
{
matches.append(index);
}
}
}
}
return matches;
}
// Dictionary merging/editing.
// literalRE:
// - true: behave like dictionary::merge, i.e. add regexps just like
// any other key.
// - false : interpret wildcard as a rule for items to be matched.
bool merge
(
const bool addNonExisting,
dictionary& thisDict,
const dictionary& mergeDict,
const bool literalRE,
const HashTable<wordList, word>& shortcuts
)
{
const wordList shortcutNames(shortcuts.toc());
bool changed = false;
// Save current (non-wildcard) keys before adding items.
HashSet<word> thisKeysSet;
{
List<keyType> keys = thisDict.keys(false);
forAll(keys, i)
{
thisKeysSet.insert(keys[i]);
}
}
// Pass 1. All literal matches
forAllConstIter(IDLList<entry>, mergeDict, mergeIter)
{
const keyType& key = mergeIter().keyword();
if (key[0] == '~')
{
const word eraseKey = key.substr(1);
if (thisDict.remove(eraseKey))
{
// Mark thisDict entry as having been match for wildcard
// handling later on.
thisKeysSet.erase(eraseKey);
}
changed = true;
}
else if (literalRE || !(key.isPattern() || shortcuts.found(key)))
{
entry* entryPtr = thisDict.lookupEntryPtr
(
key,
false, // recursive
false // patternMatch
);
if (entryPtr)
{
// Mark thisDict entry as having been match for wildcard
// handling later on.
thisKeysSet.erase(entryPtr->keyword());
if
(
addEntry
(
thisDict,
*entryPtr,
mergeIter(),
literalRE,
shortcuts
)
)
{
changed = true;
}
}
else
{
if (addNonExisting)
{
// not found - just add
thisDict.add(mergeIter().clone(thisDict).ptr());
changed = true;
}
else
{
IOWarningInFunction(mergeDict)
<< "Ignoring non-existing entry " << key
<< endl;
}
}
}
}
// Pass 2. Wildcard or shortcut matches (if any) on any non-match keys.
if (!literalRE && thisKeysSet.size() > 0)
{
// Pick up remaining dictionary entries
wordList thisKeys(thisKeysSet.toc());
forAllConstIter(IDLList<entry>, mergeDict, mergeIter)
{
const keyType& key = mergeIter().keyword();
if (key[0] == '~')
{
const word eraseKey = key.substr(1);
// List of indices into thisKeys
labelList matches
(
findMatches
(
shortcuts,
shortcutNames,
thisKeys,
eraseKey
)
);
// Remove all matches
forAll(matches, i)
{
const word& thisKey = thisKeys[matches[i]];
thisKeysSet.erase(thisKey);
}
changed = true;
}
else
{
// List of indices into thisKeys
labelList matches
(
findMatches
(
shortcuts,
shortcutNames,
thisKeys,
key
)
);
// Add all matches
forAll(matches, i)
{
const word& thisKey = thisKeys[matches[i]];
entry& thisEntry = const_cast<entry&>
(
thisDict.lookupEntry(thisKey, false, false)
);
if
(
addEntry
(
thisDict,
thisEntry,
mergeIter(),
literalRE,
HashTable<wordList, word>(0) // no shortcuts
// at deeper levels
)
)
{
changed = true;
}
}
}
}
}
return changed;
}
int main(int argc, char *argv[])
{
#include "addDictOption.H"
argList::addOption
(
"subDict",
"name",
"specify the subDict name of the replacements dictionary"
);
argList::addOption
(
"instance",
"name",
"override instance setting (default is the time name)"
);
// Add explicit time option
timeSelector::addOptions();
argList::addBoolOption
(
"literalRE",
"treat regular expressions literally (i.e., as a keyword)"
);
argList::addBoolOption
(
"enableFunctionEntries",
"enable expansion of dictionary directives - #include, #codeStream etc"
);
argList::addBoolOption
(
"disablePatchGroups",
"disable matching keys to patch groups"
);
#include "addRegionOption.H"
#include "setRootCase.H"
#include "createTime.H"
// Optionally override controlDict time with -time options
instantList times = timeSelector::selectIfPresent(runTime, args);
if (times.size() < 1)
{
FatalErrorInFunction
<< "No times selected." << exit(FatalError);
}
forAll(times, timei)
{
word instance;
if (args.optionFound("instance"))
{
if (times.size() > 1)
{
FatalErrorInFunction
<< "Multiple times selected with 'instance' option"
<< exit(FatalError);
}
args.optionLookup("instance")() >> instance;
}
else
{
runTime.setTime(times[timei], timei);
instance = runTime.timeName();
}
#include "createNamedMesh.H"
const bool literalRE = args.optionFound("literalRE");
if (literalRE)
{
Info<< "Not interpreting any regular expressions (RE)"
<< " in the changeDictionaryDict." << endl
<< "Instead they are handled as any other entry, i.e. added if"
<< " not present." << endl;
}
const bool enableEntries = args.optionFound("enableFunctionEntries");
if (enableEntries)
{
Info<< "Allowing dictionary preprocessing ('#include', '#codeStream')."
<< endl;
}
const int oldFlag = entry::disableFunctionEntries;
if (!enableEntries)
{
// By default disable dictionary expansion for fields
entry::disableFunctionEntries = 1;
}
const bool disablePatchGroups = args.optionFound("disablePatchGroups");
if (disablePatchGroups)
{
Info<< "Not interpreting any keys in the changeDictionary"
<< " as patchGroups"
<< endl;
}
fileName regionPrefix = "";
if (regionName != fvMesh::defaultRegion)
{
regionPrefix = regionName;
}
// Make sure we do not use the master-only reading since we read
// fields (different per processor) as dictionaries.
regIOobject::fileModificationChecking = regIOobject::timeStamp;
// Get the replacement rules from a dictionary
const word dictName("changeDictionaryDict");
#include "setSystemMeshDictionaryIO.H"
IOdictionary dict(dictIO);
const dictionary* replaceDictsPtr = &dict;
if (args.optionFound("subDict"))
{
word subDictName(args.optionLookup("subDict")());
replaceDictsPtr = &dict.subDict(subDictName);
}
const dictionary& replaceDicts = *replaceDictsPtr;
Info<< "Read dictionary " << dict.name()
<< " with replacements for dictionaries "
<< replaceDicts.toc() << endl;
// Always read boundary to get patch groups
// ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Info<< "Reading polyMesh/boundary file to extract patch names"
<< endl;
// Read PtrList of dictionary as dictionary.
const word oldTypeName = IOPtrList<entry>::typeName;
const_cast<word&>(IOPtrList<entry>::typeName) = word::null;
IOPtrList<entry> dictList
(
IOobject
(
"boundary",
runTime.findInstance
(
regionPrefix/polyMesh::meshSubDir,
"boundary",
IOobject::READ_IF_PRESENT
),
polyMesh::meshSubDir,
mesh,
IOobject::READ_IF_PRESENT,
IOobject::NO_WRITE,
false
)
);
const_cast<word&>(IOPtrList<entry>::typeName) = oldTypeName;
// Fake type back to what was in field
const_cast<word&>(dictList.type()) = dictList.headerClassName();
// Temporary convert to dictionary
dictionary fieldDict;
forAll(dictList, i)
{
fieldDict.add(dictList[i].keyword(), dictList[i].dict());
}
if (dictList.size())
{
Info<< "Loaded dictionary " << dictList.name()
<< " with entries " << fieldDict.toc() << endl;
}
// Extract any patchGroups information (= shortcut for set of
// patches)
HashTable<wordList, word> patchGroups;
if (!disablePatchGroups)
{
patchGroups = extractPatchGroups(fieldDict);
if (patchGroups.size())
{
Info<< "Extracted patch groups:" << endl;
wordList groups(patchGroups.sortedToc());
forAll(groups, i)
{
Info<< " group " << groups[i] << " with patches "
<< patchGroups[groups[i]] << endl;
}
}
}
// Every replacement is a dictionary name and a keyword in this
forAllConstIter(dictionary, replaceDicts, fieldIter)
{
const word& fieldName = fieldIter().keyword();
Info<< "Replacing entries in dictionary " << fieldName << endl;
// Handle 'boundary' specially:
// - is PtrList of dictionaries
// - is in polyMesh/
if (fieldName == "boundary")
{
Info<< "Special handling of " << fieldName
<< " as polyMesh/boundary file." << endl;
// Get the replacement dictionary for the field
const dictionary& replaceDict = fieldIter().dict();
Info<< "Merging entries from " << replaceDict.toc() << endl;
// Merge the replacements in. Do not add non-existing entries.
merge(false, fieldDict, replaceDict, literalRE, patchGroups);
Info<< "fieldDict:" << fieldDict << endl;
// Convert back into dictList
wordList doneKeys(dictList.size());
label nEntries = fieldDict.size();
forAll(dictList, i)
{
doneKeys[i] = dictList[i].keyword();
dictList.set
(
i,
fieldDict.lookupEntry
(
doneKeys[i],
false,
true
).clone()
);
fieldDict.remove(doneKeys[i]);
}
// Add remaining entries
label sz = dictList.size();
dictList.setSize(nEntries);
forAllConstIter(dictionary, fieldDict, iter)
{
dictList.set(sz++, iter().clone());
}
Info<< "Writing modified " << fieldName << endl;
dictList.writeObject
(
runTime.writeFormat(),
runTime.writeFormat(),
IOstream::UNCOMPRESSED,
true
);
}
else
{
// Read dictionary
// Note: disable class type checking so we can load field
Info<< "Loading dictionary " << fieldName << endl;
const word oldTypeName = IOdictionary::typeName;
const_cast<word&>(IOdictionary::typeName) = word::null;
IOobject fieldHeader
(
fieldName,
instance,
mesh,
IOobject::MUST_READ_IF_MODIFIED,
IOobject::NO_WRITE,
false
);
if (fieldHeader.typeHeaderOk<IOdictionary>(false))
{
IOdictionary fieldDict(fieldHeader);
const_cast<word&>(IOdictionary::typeName) = oldTypeName;
// Fake type back to what was in field
const_cast<word&>(fieldDict.type()) =
fieldDict.headerClassName();
Info<< "Loaded dictionary " << fieldName
<< " with entries " << fieldDict.toc() << endl;
// Get the replacement dictionary for the field
const dictionary& replaceDict = fieldIter().dict();
Info<< "Merging entries from " << replaceDict.toc() << endl;
// Merge the replacements in (allow adding)
merge(true, fieldDict, replaceDict, literalRE, patchGroups);
Info<< "Writing modified fieldDict " << fieldName << endl;
fieldDict.regIOobject::write();
}
else
{
WarningInFunction
<< "Requested field to change " << fieldName
<< " does not exist in " << fieldHeader.path() << endl;
}
}
entry::disableFunctionEntries = oldFlag;
}
}
Info<< "\nEnd\n" << endl;
return 0;
}
// ************************************************************************* //