Implement Code level Performance Analysis using the PMUv3 plugin
Introduction
PMUv3 plugin features
Download and build the PMUv3 plugin
Instrument one section of code
Plot, visualize, and analyze the results
Instrument multiple sections of code
Review
Next Steps

  Log an issue

  Fork and edit

  Discuss on Discord

What you've learned

You should now know how to:

  • Generate a fine-grained, precise measurement of functions and other sections of code.
  • Instrument your code to analyze a single section or multiple sections using the provided instrumentation scenarios.
  • Run and collect performance metrics and raw event values for any of the 15 event groups (bundles) in a single run.
  • Use a tool to plot raw PMU event values along with KPI metric values such as MPKI, stalls, and IPC to aid performance visualization.

Knowledge Check

To permanently enable userspace access for performance monitoring, you need to add kernel.perf_user_access = 1 to the /etc/sysctl.conf file.

The PMUv3 plugin can be used to analyze performance of Python applications.

What are the advantages of using the PMUv3 plugin?


Back
Next