Tango: Precise Performance Measurement through Direct Comparison
summary
Argues that running two builds side‑by‑side removes common‑mode noise and yields reproducible results in noisy cloud environments, implements the idea as an open‑source Rust library Tango that loads two binaries into one process and alternates concurrent runs, demonstrates comparing simple loops over 1M random numbers and reliably measuring about half a percent difference, and shows Paired Benchmarking (Paired Benchmarking) works across Linux, macOS and Windows.
AI-written summary · about
chapters (4)
related talks
27mImproving Rust Performance Through Profiling and Benchmarking
31mBenchmarking and Optimization of Rust Libraries
24mRust performance debugging with TUIs and LLMs
55mThe performance Addict's Toolbox
26mAllocating Less: Really Thin Rust Cloud Apps
01h 00mFrom Tiny Chips to Server Racks: Deploying AI Everywhere with Rust
28mBuild bigger in less time: code testing beyond the basics
59mParallel Programming in Rust: Techniques for Blazing Speed