> At least that way we don't need to benchmark and it would be obvious that we're then using the most efficient shader possible.
We still need to benchmark, because, at least for a scene that contains both opaque and translucent elements, we have the potential overhead of switching GL programs.
> At least that way we don't need to benchmark and it would be obvious that we're then using the most efficient shader possible.
We still need to benchmark, because, at least for a scene that contains both opaque and translucent elements, we have the potential overhead of switching GL programs.