Hi all,
I'm finishing up my MPH in epi and I'm getting ready to start my PhD, so I figured I would chime in since I've been through this already.
I would say that SAS is by far the most widely used software in epi. Some people do use R, but not nearly as widely as SAS (for a variety of legitimate and not legitimate reasons which I won't go into here). Its fine if you don't know what SAS is, you will learn it during your courses. There are also tons of online resources available if you want to get started early, as well as a free edition that you can download.
SAS can run well on a Mac with virtualization software like Virtual Box or Parallels, so dual booting is not necessary. Any current model Mac, including a Macbook Air (which I use), is sufficient. I only recommend that you have a large enough hard drive to store all of your personal things along with SAS and Windows on your virtual machine. Because of how SAS handles data, it is actually the hard drive that will allow you to work with very large datasets, not the RAM, so if you are going to use big datasets (like genetic data or large surveillance datasets), upgrade the hard drive. If you are using R, however, your RAM will be the limiting factor.
GIS software is big and resource intensive (and expensive!). You will want to have a lot of RAM and processing power if you want to run it on a laptop, so most people just use school computers.
Hope this helps!