Benchmarking Go Code

Benchmarking is a crucial practice to measure the performance of your Go code and identify bottlenecks. The Go standard library provides a testing package which is well-suited for benchmarking.

Basic Benchmark Function

Here's a simple example of a benchmark function to measure the performance of a function:

package main

import (
	"testing"
)

// Function to benchmark.
func Sum(x, y int) int {
	return x + y
}

func BenchmarkSum(b *testing.B) {
	for i := 0; i < b.N; i++ {
		_ = Sum(1, 2)
	}
}

To run the benchmark, use the command: go test -bench=.

Benchmarking with Setup and Teardown

Often you'll want to include setup or teardown logic with your benchmarks:

package main

import (
	"testing"
)

func ExpensiveSetup() int {
	// Simulate an expensive operation.
	return 42
}

func BenchmarkWithSetupTeardown(b *testing.B) {
	data := ExpensiveSetup()
	b.ResetTimer() // Reset timer after setup

	for i := 0; i < b.N; i++ {
		_ = data + i
	}

	b.StopTimer() // Stop timer before teardown
}

Benchmarking with Parallelism

You can test how well your code performs using parallel execution:

package main

import (
	"testing"
)

func ParallelTask() {
	// Simulate some work.
}

func BenchmarkParallelTask(b *testing.B) {
	b.RunParallel(func(pb *testing.PB) {
		for pb.Next() {
			ParallelTask()
		}
	})
}

Best Practices

• Use b.ResetTimer() after any setup to ensure only the execution time of the code you're benchmarking is measured.
• Use b.StopTimer() and b.StartTimer() around parts of the benchmark that shouldn't be timed, like setup and teardown.
• Be mindful of external system effects—ensure controlled conditions to get consistent results.
• Use benchmarks to explore different variations of your code for optimization opportunities.

Common Pitfalls

• Forgetting to reset the timer (b.ResetTimer()) after setup.
• Running benchmarks on a machine with varying loads can result in inconsistent results.
• Failing to ensure benchmark code does meaningful work; avoid empty loops that the compiler might optimize away.
• Not using parallel benchmarks for functions intended to be used concurrently.

Performance Tips

• Consider using -benchmem flag to report memory allocations when running benchmarks.
• Profile your code with pprof to find hotspots and memory usage patterns.
• Compare benchmarks before and after changes to evaluate their impact on performance.
• Experiment with different sizes or types of input data to see how they affect performance.
• Use the b.ReportAllocs() method to ensure memory allocations are reported.