We're only running on CUDA, and also that logic is no longer needed since we've updated the engine settings to detect vacuum simulations.