Multi-Language Queue Benchmark
Fair, reproducible measurements of in-process queue libraries across five languages.
This is a measurement lab, not a leaderboard of “the fastest queue in the world.” Compare queues inside one language and one communication category (thread, async, or opt-in process / shared memory / durable). Cross-language times are directional only.
| Go here | Why |
|---|---|
| Benchmark design | Categories, tests, how to read a result |
| Comparison rules | What we never rank together |
| Queue categories | Thread, async, process, shared, durable |
| Architecture | How the suite is built |
| Modes | smoke / all-single / full / research |
| Metrics | What the columns mean (enqueue / dequeue / handoff) |
| Adding a queue | Drop in one library |
| Adding a language | New runner contract |
| Queues 101 | Types, backpressure, lock vs lock-free |
Languages
Quick start
./scripts/check-host-requirements.sh python
./python/scripts/run-benchmarks.sh smoke