<pcwalton> is there a faster way to linearly interpolate varyings on a quad than to triangulate the quad and use barycentric coordinates on the triangle the point is in? <pcwalton> (apologies for the basic question)