SESSION I:  Introduction to the Labs and Point to Point 
                                       Communication with MPI 

• Blocking communications with MPI
  • Getting started with "Hello World"
  • Blocking Send
  • Sending in a ring

• Non-Blocking communications with MPI
  • Replace MPI_Send with MPI_Isend and MPI_Recv with MPI_Irecv

    C and FORTRAN examples

SESSION II : MPI Derived Datatypes

• Contiguous
  • Using MPI_Type_contiguous

• Vector and Hvector
  • Define and send MPI vector datatypes
  • Timing MPI vector datatypes

• Indexed and Hindexed
  • Using the MPI_Type_indexed routine to pick up some irregularly spaced elements of an Array and send them to each processor.

• Struct
  • Using MPI_Type_struct to share data

SESSION III: Collective Communication with MPI

• Data Movement
  • Broadcast
    • Broadcast
  • Gather and Scatter
    • MPI_Gather and MPI_Scatter

  • Allgather
    Using MPI_Allgather to collect parts of matrix-vector product from their respective processes to all processes

  • Alltoall
    Using MPI_Alltoall to perform matrix transpose

• Collective computation
  • MPI_Reduce
    • MPI_REDUCE
    • Finding PI using MPI collective operations

  • MPI_AllReduce
    • MPI_Allreduce

  • Synchronization
    • Collective barrier