PROFILING.md

Profiling Go Code

You can refer to the 2 following links:

• How to write benchmarks in Go
• pprof documentation

We will leverage on Go language's built-in tools for generating benchmarks.

We will then use pprof to profile the code.

1 - Install pprof tool

go install github.com/google/pprof@latest

2 - Run the provided benchmark

The code used to run the benchmark is available here

go test -v -benchmem -cpuprofile cpu.prof -bench=. ./...

3 - Launch the profiler (web interface)

pprof -http=localhost:5000 cpu.prof

This command starts a web server listening on http=localhost:5000

Feel free to use a different port number.

4 - Browse and analyze profiling data

Open http=localhost:5000 in a browser and go through execution details to identify parts of the code to optimize.

You can use pprof's VIEW menu to choose which view to display.

5 - Improve your code and repeat

Once you have a new improved version ot the code, restart from step 2.