This paper describes the “Supercomputing Simulation and Technologies” master’s degree course, taken during one’s final year in the Faculty of Computational Mathematics and Cybernetics program at Lomonosov Moscow State University. According to the curriculum, students are introduced to parallel computation early in their studies and then challenged with an in-depth understanding of how the algorithm is implemented in the hardware and how to study and improve algorithm performance. The goal is to produce graduates who specialize in the use of high-performance computing (HPC) systems in multiple fields of study.

Two practical assignments are a cornerstone of the course. The first of these assignments is to “compose a description of the structure and properties of a parallel algorithm,” an approach that the authors believe is unique to their program and also quite important in training computational scientists because the student must bring together and apply material learned in prior classes.

Students were asked to apply the AlgoWiki approach for describing an algorithm, focusing on information structure and parallelisms, and then build and analyze a data flow graph for the algorithm. The authors believe that this exercise is necessary for evaluating complexity and looking for bottlenecks. Multiple levels of detail were used to build graphs that can be examined at the meta and micro levels. Finally, students could either find an implementation of the algorithm or write their own and analyze the algorithm’s scalability. Students had to explain any peculiarities found in these scalability graphs, for example, accelerated results due to fortunate memory cache placements, and identify reasons for poor scalability.

Iteration of the reports between students and tutors led to a very strong set of final reports, and some students even continued their optimization work after the course was over. The authors believe this course was highly productive for students and encourage others to replicate their approach.

The paper is well written and contains detailed explanations when needed to help readers understand what was unique in the course approach.