ECE 563 Programming Parallel Machines -- running parallel jobs and related information
Purdue provides a scholar cluster that gives access Purdue supercomputing resources for educational purposes and we will have access to this for assignments in this class. If you have access to a distributed cluster in your job or research lab you may use that instead of the scholar machines. For timing runs, use a resource that gives your job exclusive access to the resources used to run the job. Scholar provides this.
Instructions on how to compile and run shared memory OpenMP jobs, and distributed memory MPI jobs on the Scholar system can be found in the scholar guide
Shared memory -- OpenMP, Pthreads documentation
A comprehensive OpenMP tutorial from the Lawrence Livermore National Lab (LLNL).
A shorter OpenMP tutorial from a former graduate student at Purdue.
Instructions and scripts provided by a student in a previous semester can be found at scholar cookbook.
Instructions on how to not have to enter your password a gazillion
times every time you run an mpi program is available
here.