Comment by stingraycharles

I wonder whether they have an explanation for this. Probably more overhead in the abstractions necessary for parallel processing?