Carbon Footprint–Aware Software Development
summary
Presents measurements of software energy and carbon impact using Linux tools (perf, Flame Graphs, powertop) and an example SQLite drop-table benchmark showing energy cut from 37 J to 15 J by changing compiler and optimization flags (Borland vs Visual C++ with O1/O2). Gives broader context numbers (0.2 g CO2 per Google search, ~55 g CO2/hour for video, ~625,000 lb CO2 for training a 213M-parameter transformer) and proposes adding CI/CD energy regressions and using tracing tools like Argos to record end-to-end watt-hours.
AI-written summary · about
chapters (6)
- 0:00Software carbon footprint
- 3:54Digital ecology impact
- 8:49Carbon footprint growth
- 14:01Reducing footprint methods
- 20:10Energy efficiency tools
- 27:07Q&A
related talks
59mReducing The Carbon Footprint Of Your Software Systems
55mCode that Won't Heat the Planet: Engineering Sustainable Software for a Better Future
49mDoing for Sustainability, What Open Source Did for Software
37mHow to measure your software's impact on the planet
38mGreen software: How can tech contribute to worldwide sustainability effort?
38mBetter Coding for the Climate?!
18mEngineering for Sustainability
50mResponsible computing: how about a survey of measurement tools?