Go Testing Benchmarks

Benchmarks

Measure function performance with benchmark tests.

CORRECT

func Benchmark_Fibonacci(b *testing.B) {
    for i := 0; i < b.N; i++ {
        Fibonacci(10)
    }
}

func Benchmark_Fibonacci_Cases(b *testing.B) {
    cases := []struct {
        name string
        n    int
    }{
        {name: "small", n: 10},
        {name: "medium", n: 20},
        {name: "large", n: 30},
    }

    for _, bc := range cases {
        b.Run(bc.name, func(b *testing.B) {
            for i := 0; i < b.N; i++ {
                Fibonacci(bc.n)
            }
        })
    }
}

func Benchmark_StringConcat(b *testing.B) {
    b.ResetTimer()  // Exclude setup time
    for i := 0; i < b.N; i++ {
        _ = "hello" + "world"
    }
}

Run benchmarks:

go test -bench=. -benchmem

Why:

• b.N adjusted automatically for stable timing
• Sub-benchmarks compare variations
• b.ResetTimer() excludes setup
• -benchmem shows allocation stats

WRONG

func Benchmark_Fibonacci(b *testing.B) {
    Fibonacci(10)  // Missing loop
}

func Benchmark_NoReset(b *testing.B) {
    // Expensive setup
    data := generateLargeData()
    // Missing b.ResetTimer()
    for i := 0; i < b.N; i++ {
        Process(data)
    }
}

Problems:

• No b.N loop = invalid benchmark
• Setup time included in measurement
• Inaccurate performance results