TOPIC: Sending & Receiving Data With MPI GOAL: Parallel Search Algorithm Code: send/. We will deal with the different methods for sending/receiving data in MPI (also referred to as "send modes"). We will talk about some common problems in communicating (deadlock) and how to avoid them. We will also treat the possibility of so-called non-blocking communication where processes continue their operation while the communication is still in progress. As an example we will deal with the parallel implementation of a simple search for specific data in a vector. For this purpose we will employ a "master-and-slave" strategy, i.e. one of the processors is designated the "master" who co-ordinates the tasks of the other processors.