Hi,
I'm using the move_mean in bottleneck but i see no speed up. I'm fairly sure it happens because bottleneck calls its slow implementation. Is there a way to prevent calling the slow implementation? and under which conditions is the slow implementation called?