Motivating Complexity Understanding By Profiling Energy Usage
summary
Claims that energy measurements make complexity concrete and motivating by showing real CO2 costs, and demonstrates it with a $120 setup: a Raspberry Pi 4B, DHT22 sensor, Makerhawk USB power monitor and external meter, producing sample runs that emit about ~1 gram CO2 overall. Sorting experiments show bubble sort uses far more energy than other O(n^2) sorts and moving to n log n cuts energy to roughly one fifth while scaling from 50k to millions. The team builds Draculog (Node/React/MariaDB front end, Python back end using PyRAPL) as open source and deploys a Linux/Intel server for larger runs.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:15Teaching Complexity is Hard
- 1:35What Students Care About
- 2:24The Zeitgeist
- 3:02Research Questions
- 3:43Evaluating Complexity
- 4:31Why Profiling is Less Interesting
- 6:08Experimental System
- 9:46Software
- 12:04Survey
- 12:29Limitations
related talks
31mTrack your code's CO2 emissions with Code Carbon
59mReducing The Carbon Footprint Of Your Software Systems
42mEnergy-Efficient Software Architecture for Developers
27mMeasuring Carbon Footprint of Personal Computing
56mProve going for a walk makes you more productive by measuring CO2 with an IoT device.
28mMore Data Please: Hands on Green Cloud Experiments
50mEnergy Efficient Design Principles of Software Engineering
34mCarbon Footprint–Aware Software Development