I don't have a solution to your problem, but I have also noticed some performance issues that I think are tied to the operating system. Old code I wrote ran in 23.2 minutes using Mathematica 9.0 and Windows 7. The exact same code running under Mathematica 10.3 and Windows 10 runs in 35.6 minutes. It is very frustrating especially when I spent so much time optimizing performance under Mathematica 9. I hope you make progress understanding this.
