I am trying to decide how to tune Vagrant VMs for performance. In this case, the application is web development (linux + ruby + node.js + java, etc).
I usually have to run either one VM by itself, or two VMs simultaneously. In my case, it's on a Macbook Pro, but for the organization, various types of machines will be used.
I have a Vagrantfile that dynamically selects the number of CPUs to use for each VM (e.g., half the available cores, all the available cores, etc), and can also do this with RAM (e.g., 1/4 of total system RAM).
What I'm wondering is: if one has all their Vagrant VMs set to use half or all of the CPU cores, and they run 2 or more VMs simultaneously, how does this play out?
Is it better, or worse, technically, in terms of total performance, given this application (web development)?
