I'm currently attempting to implement a skeletal animation system in OpenGL. This system is based around a bone structure with bone weights for each vertex. I am wondering what the easiest way to process the vertex data is. The bones have particular transformations applied, which should then be applied to the vertices according to their respective weights. I was considering finding the matrix for each vertex, multiplying it by the vertex position in software, and then sending it to the graphics card. However, if I have to calculate the new position of each vertex every frame, based on interpolating the bone matrices, won't that be quite a slow-down? Is there a way that will allow more processing to be done in hardware, like matrix multiplication, etc.? I'm hoping to make the system compatible with vertex arrays/VBOs.