Lecture VII

This video lecture discusses running OpenMP programs on the Bura Supercomputer. OpenMP is a multi-threaded and shared memory parallelism interface used to direct parallelism in programs. The lecture explains that OpenMP programs can be run on the cluster or SMP (Symmetric Multiprocessing) partitions of the supercomputer. The lecturer demonstrates how to compile and run OpenMP programs using examples like a Hello World program, a Pi computation program, and a Jacobi linear solver program. The presenter explains the compilation process and provides instructions on setting up and running Slurm scripts to execute the programs on the supercomputer. They also discuss the importance of defining the number of threads and allocating memory per CPU in the Slurm script. The lecture concludes by showcasing the execution results and highlighting the impact of thread count and iteration on the program’s execution time.