2D hydro code

Code for the lecture 'High Performance Computing' in spring semester of 2017.

Based on a simple 2D hydro code by

• (C) Romain Teyssier : CEA/IRFU (original F90 code)
• (C) Pierre-Francois Lavallee : IDRIS (original F90 code)
• (C) Guillaume Colin de Verdiere : CEA/DAM (for the C version)

MPI/OpenMP functionallity added by Yannick Boetzel.

How to run it

Got to Src folder and compile the code:

cd C/MonoRef/Src
make clean
make {mpi|mpiomp|omp}

Choose either one of the options in {mpi|mpiomp|omp} depending if you want to run the MPI, OpenMP or hybrid MPI/OpenMP version. mpi and mpiomp can both be used on the mpiomp branch, omp can only be used on the omp branch.

The makefile will compile everything and create a file hydro_{mpi|mpiomp|omp}. You can then run it by first specifying the number of OpenMP threads

export OMP_NUM_THREADS = 2

then run it either by

mpirun -np 4 --map-by core --bind-to core hydro_mpiomp -i input.nml

for the MPI or hybrid MPI/OpenMP version (4 cores, binds threads to cores), or by

./hydro_omp -i input.nml

for the pure OpenMP version. The input files are located in the Input folder.

You can also use one of the two python scripts weak.py and strong.py to test weak and strong scaling of the code. Usage here is

python weak.py -g 500 500 -s 100 -n 2 4 8 12 16
python strong.py -g 500 500 -s 100 -n 2 4 8 12 16

This will automatically run hydro_mpiomp on a grid of 500x500 for 100 steps using 2, 4, 8, 12 and then 16 cpu cores. If you want to enable OpenMP just run it with the flag --omp

python weak.py -g 500 500 -s 100 -n 2 4 8 12 16 --omp

An overview of all options is given with

python weak.py --help