chi_mpi_utils.cc

Go to the documentation of this file.

#include "chi_mpi_utils.h"
 
namespace chi_mpi_utils
{
 
/**Returns the current rank on the specified communicator.*/
int GetLocationID(MPI_Comm mpi_comm)
{
  int location_id;
  MPI_Comm_rank(mpi_comm, &location_id);
 
  return location_id;
}
 
/**Returns the total number of ranks on the specified communicator.*/
int GetProcessCount(MPI_Comm mpi_comm)
{
  int process_count;
  MPI_Comm_size(mpi_comm, &process_count);
 
  return process_count;
}
 
/**Given each location's local size (of items), builds a vector (dimension
* comm-size plus 1) of where each location's global indices start and end.
* Example: location i starts at extents[i] and ends at extents[i+1]*/
std::vector<uint64_t> BuildLocationExtents(uint64_t local_size, MPI_Comm comm)
{
  const int process_count = GetProcessCount(comm);
  // Get the local vector sizes per process
  std::vector<uint64_t> local_sizes(process_count, 0);
  MPI_Allgather(&local_size, // sendbuf
                1,
                MPI_UINT64_T,       // sendcount + sendtype
                local_sizes.data(), // recvbuf
                1,
                MPI_UINT64_T, // recvcount + recvtype
                comm);       // communicator
 
  // With the vector sizes per processor, now the offsets for each
  // processor can be defined using a cumulative sum per processor.
  // This allows for the determination of whether a global index is
  // locally owned or not.
  std::vector<uint64_t> extents(process_count + 1, 0);
  for (size_t locJ = 1; locJ < process_count; ++locJ)
    extents[locJ] = extents[locJ - 1] + local_sizes[locJ - 1];
  extents[process_count] = extents[process_count - 1] + local_sizes.back();
 
  return extents;
}
}