Yup runtime filtering + sorting using _eval will take up more CPU cycles than without using _eval. So best to benchmark it