Not sure what you're trying to say.
I'm asking for an apples-to-apples comparison here. Using one figure from a random CPU on github and comparing it to another figure from a different random CPU on github is not an apples to apples comparison.
I don't think anyone claimed xxhash would be 5x faster than metrohash. (Maybe you are getting the 5x number from metrohash's claim that it is 5x faster than siphash?)