Hello.
Definitely, motion capture utilizing markers outperforms any markerless system because it's a lot easier to determine actor's pose by good visible markers. Yes, our motion recognition software is far from realtime now. But it's more convenient for actor (no special suit), is less expensive, and allows tracking nearly any kind of movement. That is why you are considering it right now.
Sure, you need more powerful hardware to employ iPi Studio at a more comfortable speed. The crucial component is video card, which performs main computations. GeForce 8400GS seems outdated, you'll need more modern model. See our comparison chart of video cards:
http://wiki.ipisoft.com/Cameras_and_accessories#Video_Card. (Chart was actual on Septemper 2010, now prices has lowered, and more productive models are available at the same prices.) For example, processing of 1 Kinect frame on "ATI Radeon HD 6670 DDR5 (about $100) / Intel Core i5-2380P / 8Gb" system is about 0.7s.
With two Kinects processing time will be approximately two times greater than with one. But using two sensors provides better accuracy and allows for tracking more complex movements, such as full 360 degrees rotations.