Feature Request: Fix regression in Dynamic Resolution versus demo

[jwittlincohen]

There is an apparent regression in the way DRS was implemented in the full game versus how it operated in the demo. DRS is designed to maintain high GPU usage, in the 90%+ range, modifying the internal resolution with the selected upscaler, to achieve a given FPS target. For example, Horizon Forbidden West will make very granular internal resolution changes, maintaining 95%+ GPU utilization, while achieving a locked target FPS. In contrast, on my RTX 4090, FF16 will leave the GPU at 75% utilization when DRS is enabled, resulting in unnecessarily low internal resolution. While your mod allows forcing a higher internal resolution at the low-end of the range, this makes it less useful at actually maintaining a set frame rate. This is also a regression from the behavior in the demo, where game would target a higher GPU utilization, in the mid 80% range. ( FF16 also only allows resolution changes in 5% increments, so it's less granular than the Nixxes ports that implement DRS.)

The low GPU usage while using DRS versus the demo is well documented here (not my post): https://www.reddit.com/r/FFXVI/comments/1fl2mpi/dynamic_resolution_broken_in_the_full_game/

I was wondering if it was possible to change the game's behavior to target higher GPU utilization when DRS is enabled, as this would make DRS a lot more useful. I found DRS still can be beneficial by setting a floor of 70% of native resolution (at 4K), as this allows nearer to native 4K in less demanding scenes while still maintaining above 4K DLSS Quality at all times, but this naturally makes DRS less useful at preventing FPS drops in very demanding scenes. If the game targeted, say, 95% GPU utilization, the game could provide the full benefits of DRS, as we've seen in the Nixxes ports for Sony titles. I have no idea if this is feasible so I apologize if it's just not something you can do.