If you search online for information on how to benchmark in Rust, you will probably see a bunch of guides on how to do it in nightly Rust, but not many on how to do it in stable Rust. This is because the built-in Rust benchmarks are only available on the nightly channel. Let's start by explaining how the built-in benchmarks work, so that we can then find out how to do it in stable Rust.
First of all, let's see how to create benchmarks for a library. Imagine the following small library (code in lib.rs):
//! This library gives a function to calculate Fibonacci numbers.
/// Gives the Fibonacci sequence number for the given index.
pub fn fibonacci(n: u32) -> u32 {
if n == 0 || n == 1 {
n
} else {
fibonacci(n - 1) + fibonacci(n - 2)
}
}
/// Tests module.
#[cfg(test)]
mod tests {
use super::*;
/// Tests that the code gives the correct results.
#[test]
fn it_fibonacci() {
assert_eq!(fibonacci(0), 0);
assert_eq...