Topics in High Performance Scientific Computing is a PhD-level special topics course covering recent advances in scientific computing. The course is presented in a seminar format, where each student investigates and presents regarding a topic of personal interest.
High Performance Scientific Computing (HPSC) introduces the computing systems, software, and methods used to solve large scale problems in science and engineering. The HPSC course focuses on parallel programming using MPI, but covers SMP parallelism using OpenMP as well. Students work on several Linux clusters and a massively parallel supercomputer.
HPSC is generally offered in the spring every other year, currently on even-numbered years.
The HPSC course consists of lecture and lab sections. As a hands-on technial course, the lab section includes weekly programming assignments in FORTRAN or C that should be completed outside of class time. Programming assignments and debugging are generally performed on CSC cluster systems, such as Hemisphere, with final scalability studies executed on the Frost supercomputer.
In addition, all students must complete a large parallel programming project during the semester. Undergraduate students in CSCI 4576 may work in groups. Graduate students in CSCI 5576 must work alone on a more extensive project.
Final grades are determined by the homework assignments (50%) and the final project (50%).