First steps

• Sequential in directory seq.
• OpenMP 3.0  in directory shm.
• MPI  in directory par.
• CUDA  in directory CUDA.
• OpenACC in directory OpenACC:
• MIC (MIC-pragmas / OpenMP 4.5) in directory MIC.
• MPI+OpenMP in directory OpenACC/par*:
• MPI+CUDA in CUDA/par*.
• MPI+OpenACC in OpenACC/par*:
• MPI+MIC:

It is strongly advised to use available Performance tool as:

• Intel Advisor, including now [2017] also a rooftop analysis:
    export ADVIXE_EXPERIMENTAL=roofline
  echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope
  advixe-gui &
  See hints regarding the analysis feeback.
•