Currently, benchmarks are only done on python2 [1]. Shouldn't this be python3? Or at least not only python2?

[1] asv.conf.json has "pythons": ["2.7"],; see
https://github.com/numpy/numpy/blob/master/benchmarks/asv.conf.json#L38