go-testing-benchmarks
Benchmark patterns for performance testing
When & Why to Use This Skill
This Claude skill provides standardized benchmark patterns for Go performance testing, enabling developers to accurately measure function execution time and memory allocations. It helps in identifying performance bottlenecks and ensuring code efficiency by following Go's testing package best practices, such as proper loop implementation and timer resets.
Use Cases
- Case 1: Measuring the execution time of core algorithms to identify performance bottlenecks in high-load applications.
- Case 2: Comparing different code implementations or data structures to determine which version is more CPU and memory efficient.
- Case 3: Profiling memory allocations using the -benchmem flag to optimize garbage collection overhead in Go services.
- Case 4: Implementing sub-benchmarks to evaluate how function performance scales across different input sizes (e.g., small vs. large datasets).
| name | go-testing-benchmarks |
|---|---|
| description | Benchmark patterns for performance testing |
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