Understanding Application Performance with the POP Metrics

What you will learn

  • Why it is difficult to analyse and understand causes of poor parallel performance
  • What trace data is and why it is challenging to interpret
  • The idea behind the POP metrics
  • How high level POP metrics aid understanding of CPU parallel application performance
  • The tracing tools used in POP assessments

Prerequisites

  • Basic understanding of HPC and parallel software

Training module

Material