256 + 8 is A LOT of samples for the deferred path. I would not do that. Also, forcing those LODs will just increase your memory footprint. Especially the texture streaming. Maybe that’s why you weren’t able to render at that res anymore?
256 + 8 is A LOT of samples for the deferred path. I would not do that. Also, forcing those LODs will just increase your memory footprint. Especially the texture streaming. Maybe that’s why you weren’t able to render at that res anymore?