The Soul of Erlang and Elixir
summary
Live demos show a single BEAM OS process running thousands of 10,000 processes on one scheduler, sustaining ~10K reported successes/sec while preemptive scheduling (<1ms) prevents one CPU-bound task from stalling others. The runtime reveals busy PIDs via process_info/runtime.top, stack traces and tracing, allows process.exit(:kill) cancellation, and supports hot code reload to fix a NaN/edge-case bug and replace a brute-force sum with the O(1) formula. Cluster demo uses swarm dependency to spawn named services; BEAM claims a 130 million limit on processes.
AI-written summary · about
chapters (14)
- 0:00Start
- 0:18The Legacy of Joe Armstrong
- 1:48Introduction
- 2:47Concurrency
- 2:57Runtime Execution Context
- 3:36Concurrent Elixir
- 5:38Microservices
- 8:10Dashboard
- 11:40Process Structure
- 15:37Separate Latencies
- 21:53Process Ids
- 26:19Current Stack Trace
- 33:22Code Walkthrough
- 41:20Parting Gifts
related talks
01h 01mThe Erlang Ecosystem
51mThe Erlang VM or How I Started Loving Parallel Progamming
48mGetting to Know the Erlang Virtual Machine, the BEAM
38mDemos and Lightning Talks
10mBack to the Future
45mElixir: une potion pour mixer complexité et concurrence en temps réel
42mIt's about (run)time -- Elixir beyond the hype
48mElixir by the Bellyful