kala provides a simple API for timing sections of code with sub-nanosecond precision. It can run and time a function while preventing interference from CPU and compiler instruction reordering. It uses CPU hardware clock counters like the Time Stamp Counter (TSC) on x86-64 or the Cycle Counter Register (CCNT) on AArch64 (with an std::time::Instant fallback on other architectures).

use kala::Stopwatch;

let sw = Stopwatch::new(); // Most efficient to create once and re-use.
let nanos: f64 = sw.time(|| 2 + 3).to_nanos(); // Runs closure _once_.
println!("Function took {nanos:.2} nanos");

kala is not a benchmarking harness; it won't run a closure in a loop and report summary statistics. You need to provide that on top of kala.

See the API documentation.

Comparison With Other Libraries

• quanta
  • Does not prevent interference from CPU and compiler instruction reordering.
  • Less precise TSC frequency estimation.
• clocksource
  • Does not prevent interference from CPU and compiler instruction reordering.
  • Does not appear to use CPU hardware clocks; always goes through OS APIs.
• libcpucycles
  • Focused on counting cycles, not time[^1].
  • C library.

The are other libraries in the same space, but all of them either don't use CPU hardware clocks, don't handle CPU/compiler instruction reordering or usually both.

[^1]: CPU cycles don't map cleanly to time due to factors like dynamic frequency scaling, memory latency and others.

Contributing

Every part of this library was written by hand by a professional software engineer with decades of experience. No AI-authored code, text or images are included.

This project rejects all contributions containing AI content. There are no exceptions.