دالة parallelize

A plain loop runs on a single core while the rest sit idle. For heavy work, that wastes most of your CPU.

Parallelism means handing different loop iterations to different cores so they all work at the same time. 🧵

Mojo's parallelize helper runs your loop body across many threads for you, so you skip the manual thread plumbing.

You write a small function that does the work for one index. Mojo calls it many times, once per iteration.

The closure goes in square brackets as a compile-time parameter, not as a normal call argument. Mojo specializes on it.

The number in the parentheses is how many work items to run. That count is an ordinary runtime value.

Use num_physical_cores from the sys module to learn how many real cores you have to spread work across.

An optional second argument sets how many workers run at once. Matching it to your cores often works best.

parallelize assumes work items do not depend on each other's order. Each call should stand on its own.

Parallelism shines when each item does real work. For tiny loops, the thread overhead can cost more than it saves.

Time the loop before and after going parallel. A real benchmark tells you whether the extra threads actually helped.

You want to run a loop body across several CPU cores in Mojo.

You import parallelize, pass a per-item closure as a bracket parameter and the count in parentheses, then size workers to your cores for a real speedup. 🚀