# Is there a Performance Penalty for LOD objects with different topology?

I have a fairly complex-shaped object I've modeled, and in creating the LOD meshes, I've merged some verts, deleted some and even added some after in cases where this manipulation produced an n-gon. The UVs look fine, but these meshes have slightly different topology and vertex placement in some spots. Is this okay? What are the ramifications?

• I don't understand this question. LODs have to have different topology. "Topology" refers to the connections between vertices. If two meshes have different numbers of vertices and triangles, the topology cannot be the same. Unless you have a different definition of "topology". Can you explain what exactly it is you are concerned about? May 16, 2021 at 17:16
• Maybe I'm not using the term correctly. I always assumed it meant the way that the components (vertices and edges) were laid out. I didn't only take away edge loops, in some cases I had to add/subtract edges and merge them etc. In a case where there were two edge loops between two UV seams, I even took the original two away and added a single edge loop so that it would be squarely in the center rather than biased towards a particular direction. That means that some of the edges/loops in the lower-poly aren't present in the higher poly. Would this create a performance penalty/problem? May 16, 2021 at 22:42
• Topplogy in mathematics and computer graphics mean different things. May 17, 2021 at 5:35
• Maybe you can edit your question and clarify a bit more what exactly you did. Especially when we are talking about performance details matter a lot. For example, how is your buffer layed out? May 17, 2021 at 8:41
• Even using the mathematical definition, exploiting topological differences in LOD are routine. Consider the case of displaying a map; it's routine for lakes and islands disappear when you zoom out. I'm unclear what "performance" means here. Is this a question about real-time APIs, or data structures for ray tracing, or what? May 18, 2021 at 0:47