DLSS requires a good CPU and better matrix acceleration part inside CUDA cores to handle ML.
The problem is that Nintendo handheld has to abide by the 9W TDP. If they enable little cores to handle DLSS, then they will have to decrease GPU usage.
Another problem is to improve efficiency, they have to cut uneccessary part from GPU to achieve smaller die size. I bet that matrix accelaration would be on the chopping block first.
The problem is that Nintendo handheld has to abide by the 9W TDP. If they enable little cores to handle DLSS, then they will have to decrease GPU usage.
Another problem is to improve efficiency, they have to cut uneccessary part from GPU to achieve smaller die size. I bet that matrix accelaration would be on the chopping block first.