Energy Implications of Your Go code
summary
Edge devices spend ~65% of energy on data movement so experiments use Intel RAPL energy counters and Go tooling (go test bench, runtime/trace) to compare three PageRank implementations: a dense matrix version, a goroutine-per-vertex Pregel variant, and a sequential stream implementation. Measurements show the sequential stream runs faster and uses less energy on sparse graphs, Pregel causes heavy cross‑core communication and heap churn, and stack locality plus reduced data movement cuts power.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:36Performance vs energy in Go
- 7:13Data movement costs
- 10:57Graph processing with Pregel
- 15:30Benchmarking energy in Go
- 21:04Power tools and traces
- 24:40Energy-saving Go guidelines
- 28:49Scheduling for energy
- 30:40Go in robotics and Python
related talks
50mEnergy Efficient Design Principles of Software Engineering
31mCreate an eco-friendly world with Green Software Engineering
27mControlling Distributed Energy Resources with Edge Computing and Go
39mCreate an Eco-Friendly World with Green Software Engineering
49mThe Future is 💚: A Journey through the Energy Efficient ..
47mEnergy Proportional Edge Computing Infrastructure
16mGreen Code: Transforming DevOps for a Sustainable Future
31mHow Go is Making us Faster
