MN://WORK/language-benchmarks

OU Benchmark

A methodology-fixed Ornstein–Uhlenbeck benchmark across C, Zig, Rust, Swift, V and TypeScript on Bun, with the same algorithm and checksum in every language.

BENCHMARK / LANGUAGESData / analyticsOpen source · measured

$ gh repo view mneves75/language-benchmarks
MN://language-benchmarks — System route

$ trace "language-benchmarks"

  1. fixed method
  2. six languages
  3. timed runs
  4. ranked table

The problem

Language benchmarks usually compare different programs, so the result measures the author more than the language.

The constraint

The methodology has to be fixed across every language, and the compiler must not be allowed to delete the work as dead code.

The approach

The same Ornstein-Uhlenbeck simulation in C, Zig, Rust, Swift, V and TypeScript on Bun: Marsaglia polar Gaussian generation, Euler-Maruyama integration, and a final checksum that prevents dead-store elimination.

The outcome

A published results table with average, median, minimum and maximum times on an Apple M4 Pro, under MIT, so the numbers can be reproduced and disputed.

Next route

Open another project

The index is the wider system.

MN://palette
↑↓ navigateEsc close