Currently I have OpenGl rendering and reading in of XML documents. Using a vertex buffer I have achieve fast rendering from the GPU. See screenshots attached. FPS averages around 15,000 fps and maxes out around 20,000.

While last notice I have decided to change how I will approach the rigid body simulation. Originally I had planned to use a particle base for my simulation, reducing each body into the points that make up its space. After lengthy discussions with the mechanical engineers I work with, I have decided that this is not the way to go. The major reason is that it would require a mass-spring system between all the particles to keep the distance between each and every other particle the same. This would be too much to calculate and could lead to unrealistic oscillations in the bodies. Therefore I have begun to switch my current implementation to a traditional rigid body simulation in which I hold an orientation, position, angular velocity and linear velocity for each body. I will in addition hold transformation of the body from a local space to the world space. I will still be able to calculate torques using a matrix solver in this method. By the end of the weekend I hope to have things folding on screen.