Open dabele opened 2 years ago
I think in a first approach, we could do an OpenMPI parallelization of the advance nodes function in
https://github.com/DLR-SC/memilio/blob/main/cpp/memilio/mobility/graph_simulation.h
which should probably be parallelizable without any synchronization.
For edges, a certain syncing will be necessary but this can also be parallelized by OpenMP in a first step.
Parallelization of Simulation with OpenMP or MPI
Possiblilties:
Maybe not threadsafe: