Backstop creates new headless chrome instance for every scenario

First of all, thank you so much for creating this tool! Overall I've been really impressed with how reliable visual testing has been using backstop.

My only complaint is that it runs rather slowly. Some of this is to be expected, since it's taking visual snapshots, but I wonder if there are ways we can optimize to make it even faster?

One thing I've noticed is that it seems to be spinning up a new headless chrome instance on a new port for every scenario. I wonder if it could be faster to run only one instance for each viewport, and reuse it for all the scenarios?

I'd love to get a discussion going about this, or other things that can be done to make the tests a little lighter.