I just tested it: 1.18: BenchmarkMap-10 13388006 77.38 ns/op BenchmarkSwitch-10 50482033 23.00 ns/op Be

I just tested it: 1.18: BenchmarkMap-10 13388006 77.38 ns/op BenchmarkSwitch-10 50482033 23.00 ns/op BenchmarkSlice-10 100000000 10.90 ns/op 1.19: BenchmarkMap-10 14947908 76.78 ns/op BenchmarkSwitch-10 49444435 23.01 ns/op BenchmarkSlice-10 100000000 10.74 ns/op edit: If I understand the release notes correctly the switch optimisations are only new for large integers and strings.