This training will review the basics of MPI-1 and 2 by focusing on the changes brought by MPI-3 around one-sided RMA communications, non-blocking collective communications. The MPI/OpenMP programming model for shared memory and accelerators will also be illustrated. Practical training sessions on the Occigen supercomputer are planned. Hybrid Programming with Threads, Shared Memory, and Accelerators (MPI-2 and MPI-3)
It will be held in the premises of CINES on the afternoon of 12 June and morning of 13 June.
Registration is open from May 6, 2019 until June 4, 2019 at noon.
The number of places available is limited to 20 participants.
