User Tools

Site Tools


workshops:practicalhpc

Note: Induction Course

Please see the CHPC Induction course for new users.

Practical HPC — CHPC Winter School

This page collects all the lecture notes, video links and worksheets from the CHPC's Winter School in one place for use by anyone who wants to self-study or teach.

The Practical HPC course is intended to teach practical skills for using the CHPC supercomputer cluster effectively and efficiently. It cover two main topics:

  1. Workflow to Job Script
  2. Scaling

over two weeks. These topics are approximately 6 sections, with a core of 4 sections and 2 optional sections.

Prerequisites

Students will need to be familiar with both the Linux command line and the concepts of High Performance Computing.

Supercomputing MOOC

This course is presented in conjunction with PRACE and EPCC who provide a free online course on the basic concepts of HPC. This provides the main background knowledge you will need to understand HPC and how the CHPC cluster works.

The first 4 weeks of the PRACE Supercomputing MOOC provide the necessary foundational information and concepts for this Practical HPC course.

You will need to complete all of the first 4 weeks of the Supercomputing MOOC to be fully prepared for all the topics in the CHPC Practical HPC course.

Linux Shell Command Line

There are many resources on the Internet that introduce the basic commands of the Linux shell. We recommend:

Welcome and Overview

  • Slides from opening session.

Introduction

The Centre

Video: CHPC Introduction part 1

https://www.youtube.com/watch?v=hyhubx-oVZI

Video: CHPC Introduction part 2

https://www.youtube.com/watch?v=B7EmuL9s_90

Video: How to Access the CHPC

CHPC Hardware

Workflow to Job Script

First we cover the basics of the user's view of the cluster and how to write, submit and monitor a job script.

Connection to the CHPC

The CHPC Cluster Environment

PBSPro Scheduler and Jobs

Scaling

Scaling is the performance of a parallel program across different number of nodes and cores. First we cover the basics of compiling a program for the cluster, then we look at performance measurement and scaling.

Compilers and Libraries

Performance

Further Reading

Parallel Programming

We are often asked about porting or writing programs to parallel hardware. This is a large and more advanced topic and so we can point you to some resources that can help you start.

Parallel Programming for Scientists and Engineers

Victor Eijkhout of TACC has written a series of superlative textbooks which are all available for free at The Art of HPC.

/app/dokuwiki/data/pages/workshops/practicalhpc.txt · Last modified: 2024/08/22 10:35 by wikiadmin