benchmark.zig

This is a tiny microbenchmark library for Zig, designed to be easy to use and very small.

Usage

You can either manually copy benchmark.zig into your project and reference it with @import("benchmark.zig"), or add it as a dependency using the Zig package manager and reference as below.

A basic usage example is shown below:

const std = @import("std");
const benchmark = @import("benchmark");
pub const main = benchmark.main(.{}, struct {
    // Benchmarks are just public functions
    pub fn arrayListWriter(b: *benchmark.B) !void {
        // Setup is not timed
        var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
        defer arena.deinit();

        while (b.step()) { // Number of iterations is automatically adjusted for accurate timing
            defer _ = arena.reset(.retain_capacity);

            var a = std.ArrayList(u8).init(arena.allocator());
            try a.writer().print("Hello, {s}!", .{"world"});

            // `use` is a helper that calls `std.mem.doNotOptimizeAway`
            b.use(a.items);
        }
    }

    pub fn allocPrint(b: *benchmark.B) !void {
        var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
        defer arena.deinit();
        while (b.step()) {
            defer _ = arena.reset(.retain_capacity);

            const result = std.fmt.allocPrint(arena.allocator(), "Hello, {s}!", .{"world"});
            b.use(result);
        }
    }
});

Running this with zig build run -Doptimize=ReleaseFast produces output similar to the following:

BENCHMARK                      ITERATIONS    TIME
arrayListWriter                  36171528    27ns/op (1s total)
allocPrint                       48251012    20ns/op (1s total)