zhyang-dev/CFDEMcoupling-PFM
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

MPI: (rank < size) is always true

Browse Source
This commit is contained in:
dqueteschiner
2015-03-31 15:48:16 +02:00
parent 851deae90b
commit ededde9190
1 changed files with 39 additions and 48 deletions
Download Patch File Download Diff File Expand all files Collapse all files

87
src/lagrangian/cfdemParticle/subModels/probeModel/particleProbe/particleProbe.C
View File

@ -120,68 +120,59 @@ void particleProbe::setOutputFile() const
void particleProbe::initialize(word typeName, word logFileName) const
{
//update the list of items to be sampled
itemCounter_ += 1;
itemsToSample_.append(logFileName);
//update the list of items to be sampled
++itemCounter_;
itemsToSample_.append(logFileName);
// init environment
//propsDict_ = particleCloud_.couplingProperties().subDict(typeName + "Props");
name_ = typeName;
const char* fileNameOut_ = logFileName.c_str();
// init environment
//propsDict_ = particleCloud_.couplingProperties().subDict(typeName + "Props");
name_ = typeName;
if(verboseToFile_)
{
if (verboseToFile_)
{
Info << "Will sample these particle IDs: " << particleIDsToSample_ << " every " << printEvery_ << endl;
Info << "Will sample these particle IDs: " << particleIDsToSample_ << " every " << printEvery_ << endl;
//initialize the output files
MPI_Comm_rank(MPI_COMM_WORLD, &rank_);
//initialize the output files
int myrank_(-1);
int numprocs_(-1);
MPI_Comm_rank(MPI_COMM_WORLD,&myrank_);
MPI_Comm_size(MPI_COMM_WORLD, &numprocs_);
rank_=myrank_;
//open a separate file for each processor
char* filecurrent_ = new char[strlen(logFileName.c_str()) + 4]; //reserve 4 chars for processor name
sprintf(filecurrent_,"%s%s%d", logFileName.c_str(), ".", rank_);
//open a separate file for each processor
char* filecurrent_;
filecurrent_= new char[strlen(fileNameOut_) + 4]; //reserve 4 chars for processor name
if (myrank_ < numprocs_)
sprintf(filecurrent_,"%s%s%d", fileNameOut_, ".", myrank_);
else //open one file for proc 0
sprintf(filecurrent_,"%s", fileNameOut_);
word file_(filecurrent_);
Info << "particleProbe for model " << name_ << " will write to file " << file_ << endl;
Info << "particleProbe for model " << name_ << " will write to file " << filecurrent_ << endl;
//generate the file streams
fileName probeSubDir = dirName_;
if (particleCloud_.mesh().name() != polyMesh::defaultRegion)
{
probeSubDir = probeSubDir/particleCloud_.mesh().name();
}
probeSubDir = probeSubDir/particleCloud_.mesh().time().timeName();
//generate the file streams
fileName probeSubDir = dirName_;
if (particleCloud_.mesh().name() != polyMesh::defaultRegion)
{
probeSubDir = probeSubDir/particleCloud_.mesh().name();
}
probeSubDir = probeSubDir/particleCloud_.mesh().time().timeName();
fileName probeDir_;
if (Pstream::parRun())
{
// Put in undecomposed case
// (Note: gives problems for distributed data running)
probeDir_ = particleCloud_.mesh().time().path()/".."/probeSubDir;
fileName probeDir_;
if (Pstream::parRun())
{
// Put in undecomposed case
// (Note: gives problems for distributed data running)
probeDir_ = particleCloud_.mesh().time().path()/".."/probeSubDir;
}
else
{
probeDir_ = particleCloud_.mesh().time().path()/probeSubDir;
probeDir_ = particleCloud_.mesh().time().path()/probeSubDir;
}
//manage files and OFstreams
mkDir(probeDir_);
sPtr = new OFstream(probeDir_/file_);
sPtrList_.append(sPtr);
//manage files and OFstreams
mkDir(probeDir_);
sPtr = new OFstream(probeDir_/filecurrent_);
sPtrList_.append(sPtr);
//Clear the containers for the fields to be probed
scalarFields_.clear();
vectorFields_.clear();
}
return;
delete [] filecurrent_;
//Clear the containers for the fields to be probed
scalarFields_.clear();
vectorFields_.clear();
}
return;
}
void particleProbe::writeHeader() const