Context Navigation

Offloading_hybrid_apps

-                      v23
+                      v24
 group. Thus, each MPI process is in charge of computing the forces or updating the particles
 of a consecutive chunk of blocks.
+This computation pattern is repeated during multiple timesteps. The communication pattern
+in each timestep consists in the fact that GPU processes exchange their particles with each
+other in a circular manner to compute the forces between their particles against the particles
+of all other GPU processes. For simplifying this pattern, the application uses a different MPI
+communicator for this circular exchanging. Once a GPU process finishes the computation of its
+forces, it sends the forces to the corresponding CPU processes, and then, it receives the
+updated particles. We perform separate sends/receives for each block.
 The available versions are: