Hello @codemonkkkey and welcome to the NVIDIA developer forums!
I saw that you added a couple of posts to the forums over the last couple of weeks, I’ll make sure to at least give some replies if I can, or get some more experts involved.
Starting of with this one it is pretty easy. If you look inside the actual technology of DLSS you will see that de-noising is an essential part of the overall deep learning model of DLSS.
But if you would want to only denoise and not upscale, you might rather consider NRD in terms of performance.
I recommend this GDC session about DLSS for more information: NVIDIA DLSS Overview & Game Integrations | NVIDIA On-Demand
I hope this helps!