AlgoBench

AlgoBench is a web app (with optional CLI) that lets developers and students run reproducible micro-benchmarks and complexity experiments across languages and machines. You define an input generator, a function under test, and a measurement plan; AlgoBench executes runs in isolated containers, captures timing, memory, and variance, and produces plots plus a short, shareable report link. It’s designed for the real world: warm-up runs, CPU pinning options, noisy-neighbor detection, and clear warnings when results are statistically weak. An AI assistant can suggest better benchmark design (e.g., sample sizes, input distributions, avoiding dead-code elimination) and explain why two results differ, but the core value is trustworthy measurement, not “AI magic.” This is useful for performance regressions, comparing libraries, teaching algorithm analysis, and making engineering decisions without hand-wavy claims.