README.md

Step 8. Writing the first benchmark: why do we need it

As you can guess, the method you implemented is not quite effective. And in order to make sure of this, we'll write a benchmark in storage/storage_test.go

func BenchmarkNearest(b *testing.B) {
	s := New()
	for i := 0; i < 100; i++ {
		s.Set(i, &Driver{
			ID: i,
			LastLocation: Location{
				Lat: float64(i),
				Lon: float64(i),
			},
		})
	}
	for i := 0; i < b.N; i++ {
		s.Nearest(1000, 123, 123)
	}
}

We'll check the work on 100, 1000, 10,000 elements in the storage

cd storage
go test -bench=.

For 100 elements

BenchmarkNearest-4         50000             24002 ns/op
PASS
ok      github.com/maddevsio/gocodelabru/step09/storage 1.460s

For 1000 elements

BenchmarkNearest-4          5000            272552 ns/op
PASS
ok      github.com/maddevsio/gocodelabru/step09/storage 1.402s

For 10000 elements

BenchmarkNearest-4           500           2799431 ns/op
PASS
ok      github.com/maddevsio/gocodelabru/step09/storage 1.714s

We can make a conclusion that the more elements in the storage, the longer we are looking for the nearest drivers.

Congratulations!

You now know how to write benchmarks. In the next step we will optimize the work of the method of issuing the nearest drivers