Harnessing Spare Cores to Breeze Through Cloud Compute
summary
Spare Cores is a Python open‑source ecosystem that inventories cloud offerings, runs hardware discovery and benchmarks (stress‑ng, Geekbench 6, IO/compression tests), and publishes a downloadable SQLite DB and APIs to pick instances by actual performance and live Spot/On‑Demand prices. It already tracks 1000+ server types, collects 100,000+ metrics, and ships a runner (Pulumi templates + CLI/SDKs) to launch instances automatically; a hosted SaaS option runs jobs for you and the dataset is updated every 5 minutes.
AI-written summary · about
chapters (9)
- 0:00Introduction and motivation
- 3:54Cloud cost inefficiencies and bin packing
- 5:43Introducing Spare Cores project
- 8:13Hardware discovery and benchmarks
- 10:24Database and data challenges
- 15:50Inspector and benchmarking tools
- 17:35Runner and Pulumi integration
- 18:38API and public data access
- 20:50Future plans and acknowledgments
related talks
17mOptimizing Cloud Compute Resources with Spare Cores
36mIntroduction to High Performance Computing on Google Cloud Platform
03mCloudcost Exporter
42mBuilding a green supercomputer in the cloud
47mCloud Native and Sustainable, Reproducible Scientific Computing
39mWhat would you do with 10^6 cores? High Performance Computing in the AWS Cloud
20mHow the cloud changed the way I do analytics and research
16mSky Computing: Charting a New Course in AI Cloud Infrastructure