A Beginner's Guide to High-Performance Computing
By Rubin H. Landau
Oregon State University
This module presents some of the general ideas behind and basic principles of high-performance computing (HPC) as performed on a supercomputer. These concepts should remain valid even as the technical specifications of the latest machines continually change. Although this material is aimed at HPC supercomputers, if history be a guide, present HPC hardware and software become desktop machines in less than a decade.
Upon completion of this module, students should be able to:
- Understand in a general sense the architecture of high performance computers.
- Understand how the architecture of high performance computers affects the speed of programs run on HPCs.
- Understand how memory access affects the speed of HPC programs.
- Understand Amdahl's law for parallel and serial computing.
- Understand the importance of communication overhead in high performance computing.
- Understand some of the general types of parallel computers.
- Understand how different types of problems are best suited for different types of parallel computers.
- Understand some of the practical aspects of message passing on MIMD machines.
The documents can be downloaded below.
Resources:
Module Document : The module document for this module.
Code Archive : The codes for this module.
Beginner's Guide to HPC Slides : A zip archive of slides to accompany this module in both pdf and pptx format.