HPC Introduction 03/2019

Introduction to HPC at the University Oldenburg

Description

This 4-day course is aimed at HPC beginners as well as experienced users who wish to learn about HPC and the use of the local clusters CARL and EDDY. This time the course includes an extra day to cover the Intel tools available on the cluster. The course will cover the following topics (extra day in bold):

• General Introduction to HPC and the Use of CARL and EDDY
• Introduction to the Job Scheduler SLURM
• Programming in an HPC Environment
• Using Intel Tools for HPC
• Introduction to Parallel Programming (MPI, OpenMP, OpenACC)
• Introduction to Matlab Distributed Compute Server (MDCS)

The course will consist of lectures and hands-on practical part. Special topics may be covered if requested.

Date, Time and Registration

The course will be held by Stefan Harfst and Wilke Trei. Atanas Atanasov from Intel will cover the second day of the course devoted to the Intel HPC tools.

• Date: March 25 – 28, 2019
• Time: 10:00 to 17:00 (watch out for announcements)
• Room: W1 0-008

The schedule for the course can be found here. Watch out for updates, in particular on the second day.

We kindly ask you to register for the course in Stud.IP (look for course No. 5.04.1036 in WS18/19 in case the link does not work). If you have problems please send us an e-mail.

Lecture Slides and Materials

Day 1 - Monday, March 25th

• Session 1: Introduction to High-Performance Computing (10:15-11:15am)
  • lecture slides [pdf]
• Session 2: Basic HPC Clusters Usage (11:30-13:00pm)
  • lecture slides [pdf]
  • example: SLURM HelloWorld [tgz]
• Session 3: HPC Cluster Environment (13:45-17:00pm)
  • lecture slides [pdf]
  • example: RandomWalk [tgz]
  • example: Matrix-Matrix Multiplication [tgz]
  • example: Array Job [tgz]

Day 2 - Tuesday, March 26th

slides will follow

Day 3 - Wednesday, March 27th

• Session 4: Introduction to Parallel Programming (10:15-11:15am)
  • lecture slides [pdf]
  • example: MPI example (Heat) [tgz]
• Session 5: Introduction to OpenMP (11:30-15:00pm)
  • lecture slides [pdf]
  • example: Pi Calculation Exercise [tgz]
  • example: Pi Calculation Exercise Solution [tgz]
  • example: Saxpy Exercise [tgz]
• Session 6: Performance Optimization (15:15-16:45pm)
  • lecture slides [pdf]
  • example: Stencil code [tgz]

Day 4 - Thursday, March 28th

• Session 7: Introduction to MDCS (10:15-16:00pm)
  • lecture slides [pdf]
  • SPMD noise example [zip]
  • SPMD contrast example [zip]
  • heat exercise [zip]