After Fitz reviewed what the students learned the day before, Sam led them in an activity designed to illustrate the advantages and challenges of parallel computing. The students separated into two groups and sat at tables to represent the different styles of parallel computation. One two-person group sat at a table to represent processes that share information (shared memory), and the other group's three people were spread out across the classroom, representing processes that must explicitly send information between each other (distributed memory). Each group's task was to solve a fifty-piece puzzle and keep track of their communications (every time they spoke with each other) and computations (every time they connected two pieces). After the activity, the students reviewed what they'd learned and discussed how it pertained to parallel programming.
Next, Fitz led the students through the basics of Message-Passing Interface (MPI) programming for distributed memory parallelism. Students learned how to send "messages" between processes, allowing them to communicate with each other. With this information, the students programmed MPI versions of their Riemann sum programs and tested them on the BCCD.