distribution_concatenation.cc File Reference

#include "generic.h"

Functions
int	main (int argc, char *argv[])

Function Documentation

main()

int main	(	int argc	,
		char *	argv[]
	)

Driver code: Testing LinearAlgebraDistributionHelpers::concatenate(...)

We concatenate three uniformly distributed distributions with nrow = 7, 5, and 3 respectively.

On one core: dist_0: nrow = 7 nrow_local rank0 = 7

dist_1: nrow = 5 nrow_local rank0 = 5

dist_2: nrow = 3 nrow_local rank0 = 3

The result distribution should have: dist_r: nrow = 15 nrow_local rank0 = 15

////////////////////////////////////

On two cores: dist_0: nrow = 7 nrow_local rank0 = 3 nrow_local rank1 = 4

dist_1: nrow = 5 nrow_local rank0 = 2 nrow_local rank1 = 3

dist_2: nrow = 3 nrow_local rank0 = 1 nrow_local rank1 = 2

The result distribution should have: dist_r: nrow = 15 nrow_local rank0 = 6 nrow_local rank1 = 9

////////////////////////////////////

On three cores: dist_0: nrow = 7 nrow_local rank0 = 2 nrow_local rank1 = 2 nrow_local rank2 = 3

dist_1: nrow = 5 nrow_local rank0 = 1 nrow_local rank1 = 2 nrow_local rank2 = 2

dist_2: nrow = 3 nrow_local rank0 = 1 nrow_local rank1 = 1 nrow_local rank2 = 1

The result distribution should have: dist_r: nrow = 15 nrow_local rank0 = 4 nrow_local rank1 = 5 nrow_local rank2 = 6

////////////////////////////////////

On four cores: dist_0: nrow = 7 nrow_local rank0 = 1 nrow_local rank1 = 2 nrow_local rank2 = 2 nrow_local rank2 = 2

dist_1: nrow = 5 nrow_local rank0 = 1 nrow_local rank1 = 1 nrow_local rank1 = 1 nrow_local rank1 = 2

dist_2: nrow = 3 nrow_local rank0 = 0 nrow_local rank1 = 1 nrow_local rank1 = 1 nrow_local rank1 = 1

The result distribution should have: dist_r: nrow = 15 nrow_local rank0 = 2 nrow_local rank1 = 4 nrow_local rank1 = 4 nrow_local rank1 = 5

The script validate.sh should run this test on 1, 2, 3 and4 cores.

{
#ifdef OOMPH_HAS_MPI
  // Initialise MPI
  MPI_Helpers::init(argc,argv);
#endif
  
  // Get the global oomph-lib communicator 
  const OomphCommunicator* const comm_pt = MPI_Helpers::communicator_pt();
 
  // How many distributions do we want to generate?
  unsigned ndistributions = 3;
 
  // What are the lengths of the above distributions?
  unsigned distlengtharray[] = {7,5,3};
 
  // The distributions to concatenate.
  Vector<LinearAlgebraDistribution> dist_to_cat;
  bool distributed = true;
 
  // Create the distribution and get the pointers to them.
  // This may be a bit long winded but we do not want to create pointers
  // to distributions using "new", then we will have to remember to call
  // delete and null the pointers. This is considered bad practice, memory
  // management should be done automatically with smart pointers.
  // We could use smart pointers but it is only available in C++11,
  // or use boost smart pointers, but we do not have boost...
  // /rant.
  for (unsigned dist_i = 0; dist_i < ndistributions; dist_i++) 
  {
    // Create the distribution.
    dist_to_cat.push_back(LinearAlgebraDistribution(
          comm_pt,
          distlengtharray[dist_i],
          distributed));
  }
  
  // The pointers to distributions to concatenate.
  Vector<LinearAlgebraDistribution*> dist_to_cat_pt;
  for (unsigned dist_i = 0; dist_i < ndistributions; dist_i++) 
  {
    dist_to_cat_pt.push_back(&dist_to_cat[dist_i]);
  }
 
  // The result distribution.
  LinearAlgebraDistribution result_distribution;
  
  // Call the concatenate function.
  LinearAlgebraDistributionHelpers::concatenate(dist_to_cat_pt,
                                                result_distribution);
  
  // Output data about the result distribution:
  // nrow()
  // first_row()
  // nrow_local()
  // distributed()
  unsigned my_rank = comm_pt->my_rank();
  unsigned nproc = comm_pt->nproc();
  std::ostringstream result_stream;
  result_stream << "out_NP" << nproc << "R" << my_rank;
  
  std::ofstream result_file;
  result_file.open(result_stream.str().c_str());
  result_file << result_distribution.nrow() << "\n";
  result_file << result_distribution.first_row() << "\n";
  result_file << result_distribution.nrow_local() << "\n";
  result_file << result_distribution.distributed() << "\n";
  result_file.close();
  
#ifdef OOMPH_HAS_MPI
  // finalize MPI
  MPI_Helpers::finalize();
#endif
  return(EXIT_SUCCESS);
} // end_of_main

References oomph::MPI_Helpers::communicator_pt(), oomph::LinearAlgebraDistributionHelpers::concatenate(), oomph::LinearAlgebraDistribution::distributed(), oomph::MPI_Helpers::finalize(), oomph::LinearAlgebraDistribution::first_row(), oomph::MPI_Helpers::init(), oomph::OomphCommunicator::my_rank(), oomph::OomphCommunicator::nproc(), oomph::LinearAlgebraDistribution::nrow(), and oomph::LinearAlgebraDistribution::nrow_local().