I don't know the mathematical answers to those questions, but I just helped run a tournament last weekend (the NFL tournament UA runs every spring semester), and the practical answers are kind of fresh in my mind.
- With the advent of computer programs that powermatch in seconds (maybe a minute if you have to do any amount of manual manipulating) is there really any logistical or practical burden to powermatching
Yes. It's not even numerical or anything--it's the simple fact that without power matching, you can literally have every round preset before the tournament even starts. If you're power matching, you have to have every ballot back before you can enter the data into the computer. So there's a logistical burden there.
- Should anyone powermatch 3 rounds given that there would not be an even distribution of aff and neg in powermatched rounds
You can. If you have an odd number of prelims, you probably should (give everyone a random aff and neg). It's really more a normative question than anything else, I'd think.
When thinking about powermatching is there any merit in thinking about what is going on at the bottom of the bracket. In other words teams getting in matches where they are evenly matched and have a chance of winning as well as reducing the number of 0-6 teams to only a couple (maybe one or none)
This is a normative question, not a mathematical question. As such, it's really up to you. This is why I don't like large schools, in that they tend to screw up the bracket by forcing pull-up rounds (3-1 vs. 4-0, for example) and things like that. But ideally, the same process at the top of the bracket should work at the bottom of the bracket, preventing both ridiculously easy wins and ridiculously harsh matches. This gives teams with decent talent but relatively little experience the best, most reasonable chance, IMO. It also means that you don't thoroughly discourage novices who take up the challenge of entering in the varsity division (always a good choice, IMO again).