Re-thinking CI bottlenecks with Distributed Execution
summary
Shows a working CI setup that uses Nx Cloud and GitHub Actions to split tasks across agents, with a 30‑line reusable workflow generator that launches orchestrator + agents. Demonstrates distributed execution by using a phone-based math demo and moving real build/lint/test tasks so 10 agents finish CI in 4m17s while 6 agents raise minimum utilization to 70% and save ~37% cost, proving distributed caching and affected-based scheduling actually reduce wall time and money.
AI-written summary · about
chapters (8)
- 0:00The CI bottleneck challenge
- 1:49Scaling with Nx and caching
- 5:16Optimizing task dependencies
- 8:25Advanced caching strategies
- 10:05Demonstrating distributed power
- 14:39Distributed task execution
- 17:53Implementing DTE with Nx
- 23:46Balancing cost and performance
related talks
19mInto the (Nx) Clouds
24mUnlocking CI Performance in the Age of Agentic Development
48mCICD with Github Actions
04h 00mNx Conf Lite 2022
55mBridging the Gap Between Ops and Developers with CI/CD
20mSpeed up your CI & Save the Environment
29mDon’t let compilation set the length of your coffee break
19mCI/CD Is Dead, Agents Need Continuous Compute and Computers