Benchmarking Go code with testing.B

Ad-hoc timing with time.Now() is inaccurate for micro-benchmarks because it does not account for loop overhead, compiler optimisations, or GC pauses.

Use the built-in benchmarking framework:

func BenchmarkMyFunc(b *testing.B) {
    // Setup outside the loop
    data := generateTestData()

    b.ResetTimer() // reset timer after setup
    for i := 0; i < b.N; i++ {
        MyFunc(data)
    }
}

// Run
go test -bench=. -benchmem ./...
// -benchmem shows allocations per op
// -benchtime=5s runs for 5 seconds instead of default 1s
// -count=3 runs benchmark 3 times for stability

Prevent dead-code elimination:

var result int
func BenchmarkMyFunc(b *testing.B) {
    var r int
    for i := 0; i < b.N; i++ {
        r = MyFunc(i)
    }
    result = r // assign to package-level var
}

testing.B runs the function b.N times, adjusting N until results are statistically stable. -benchmem instruments the allocator to report bytes/op and allocs/op, critical for spotting unnecessary heap allocations.