A method and apparatus for rendering a plurality of objects. The method includes receiving ray differential information associated with a ray, determining a first ray differential based on at least information associated with the ray and a first surface patch, processing information associated with the first ray differential and the first surface patch, determining a first resolution based on at least information associated with the first ray differential and the first surface patch, and selecting a second resolution from the plurality of predetermined resolutions. Additionally, the method includes obtaining a first geometry representation associated with the second resolution. Moreover, the method includes if the second resolution is finer than the first resolution, selecting a third resolution from the plurality of predetermined resolutions, obtaining a second geometry representation associated with the third resolution, generating a third geometry representation, and determining whether the ray and the first surface patch intersect.