เครื่องมือจับเวลาและทดสอบประสิทธิภาพ

Never guess where time goes. You first measure your code, because the slow part is rarely the part you expected. ⏱️

Mojo ships a time module with helpers to read the clock. It is your simplest tool for timing a block of work.

The now function returns the current time in nanoseconds. Capture it just before and just after the work you want to measure.

Subtract start from end to get the elapsed nanoseconds. Divide to convert into milliseconds or seconds for readable output.

Wrap only the code you care about. If you also time setup or printing, that noise hides the real compute cost you are chasing.

For serious numbers, Mojo offers a benchmark module that runs your code many times and reports a stable average automatically.

The keep function tells the compiler a value really matters. Without it, dead-code removal may delete the work you meant to time.

One run is noisy from caches and the OS. Running many iterations and averaging gives a number you can actually trust.

The first run is often slow due to cold caches and startup. A short warmup you discard makes later timings reflect steady-state speed.

Tiny inputs finish too fast to measure well. Choose an input big enough that the work takes a steady, measurable amount of time.

Always benchmark on the same machine with little running in the background. Stable conditions keep one run comparable to the next.

Think about why we time code at all.

You learned to measure first: read now from the time module, use benchmark with keep, warm up, repeat, and time steady work. 🎯