sunfluidh:sunfluidh_tuto1_parallel_config
Différences
Ci-dessous, les différences entre deux révisions de la page.
Prochaine révision | Révision précédente | ||
sunfluidh:sunfluidh_tuto1_parallel_config [2017/09/27 09:26] – créée yann | sunfluidh:sunfluidh_tuto1_parallel_config [2017/09/27 09:29] (Version actuelle) – [For an OpenMP parallel computing] yann | ||
---|---|---|---|
Ligne 11: | Ligne 11: | ||
- | === For an OpenMP parallel computing === | + | ==== For an OpenMP parallel computing |
The variable " | The variable " | ||
Ligne 29: | Ligne 29: | ||
<note important> | <note important> | ||
- | In this case, the code automatically split the grid in different | + | In this case, the code automatically split the loops associated to the computational |
</ | </ | ||
- | === For a MPI parallel computing === | + | ==== For a MPI parallel computing |
We must set the number of MPI processes used and describe how each MPI process is linked to a specific subdomain. | We must set the number of MPI processes used and describe how each MPI process is linked to a specific subdomain. | ||
Ligne 43: | Ligne 43: | ||
</ | </ | ||
- | == Example for a MPI cartesian topology == | + | === Example for a MPI cartesian topology |
We here consider 4 MPI processes, 2 along each direction. The cell size is the same as previously reported. | We here consider 4 MPI processes, 2 along each direction. The cell size is the same as previously reported. | ||
Ligne 66: | Ligne 66: | ||
- | == Example for a MPI graph topology == | + | === Example for a MPI graph topology |
We here consider 48 MPI processes, 12 along the I-direction and 4 along the J-direction. The cell size is the same as previously reported. \\ | We here consider 48 MPI processes, 12 along the I-direction and 4 along the J-direction. The cell size is the same as previously reported. \\ |
sunfluidh/sunfluidh_tuto1_parallel_config.1506497204.txt.gz · Dernière modification : 2017/09/27 09:26 de yann