Struct Benchmark 

pub struct Benchmark { /* private fields */ }

A single benchmark run.

Collects per-iteration duration samples. Call Benchmark::finish to produce a BenchmarkResult.

Example

use dev_bench::Benchmark;

let mut b = Benchmark::new("noop");
for _ in 0..10 {
    b.iter(|| std::hint::black_box(42));
}
let r = b.finish();
assert_eq!(r.samples.len(), 10);

Implementations

impl Benchmark

pub fn new(name: impl Into<String>) -> Self

Begin a new benchmark with a stable name.

pub fn iter<F, R>(&mut self, f: F) -> R

where F: FnOnce() -> R,

Run one iteration of the benchmark, capturing the duration.

Each call records exactly one sample.

Example

use dev_bench::Benchmark;

let mut b = Benchmark::new("noop");
b.iter(|| std::hint::black_box(1 + 1));
let r = b.finish();
assert_eq!(r.samples.len(), 1);

pub fn iter_with_count<F>(&mut self, n: u64, f: F)

where F: FnMut(),

Run a closure n times and record ONE sample for the entire batch.

Use for sub-microsecond operations where per-iteration timing would be dominated by Instant::now() overhead. The reported per-iteration mean is batch_duration / n.

Example

use dev_bench::Benchmark;

let mut b = Benchmark::new("hot");
b.iter_with_count(1000, || {
    std::hint::black_box(40 + 2);
});
let r = b.finish();
assert_eq!(r.samples.len(), 1);
assert_eq!(r.iterations_recorded, 1000);

pub fn finish(self) -> BenchmarkResult

Finalize the benchmark and produce a BenchmarkResult.