Techniques that facilitate rendering of a scene of arbitrary complexity without having to load all the models involved in the scene simultaneously in computer memory. A representation of the inputs and outputs of the various models involved in the scene and the dependencies between the inputs and the outputs is provided and used to facilitate the rendering process. A tessellator may tessellate the models without having to load information for all the models concurrently into computer memory. The tessellated scene may then be rendered without having to load all the models in the scene concurrently into computer memory.