Build issue with MPI and no OpenMP about euler_kokkos

[pkestene]

Hi @sampollard

thanks for reporting. Indeed I recently updated kokkos submodule and its API slightly changed, and as I was testing with USE_MPI=OFF, I forgot to fully update euler_kokkos to use the new API.
It should be fixed now. Please pull and try again.

[pkestene]

BTW, the minapp used in the article you mentioned is another miniapp but very very similar, we are solving the exact same equations (Euler equations for compressible hydrodynamics) with the same kind of numerical methods (finite volume discretization, Riemann solver, ...), but their focus is numerical precision (maybe not parallelism). Here my focus is more illustrating portability and parallelism using the Kokkos library, but surely the question of numerical precision in a parallel execution context is interesting.

[sampollard]

Fixed! I was able to compile and run a test program using

mpirun -np 6 src/euler_kokkos test/io/test_io_2d.ini