that doesn't address the bottleneck of the scripting engine

As I understand, the issue with using larger uGrids just has more to do with event-loading in the game. For example, scripts for 'random encounters' and possibly some other 'story-essential' events can be triggered from a much greater distance and subsequently 'missed' by the player, causing issue with stability and save game corruptions and such. With the standard uGrids settings, all events are triggered at their intended proximity from the player.

Now, I concede this is all hearsay, and may or may not be true at all. It is just something that I have read and SOUNDS reasonable / plausible. Please take a grain of salt with the above comment

Guys, could it be SweetFX/ReShade running for all players who suffer from low fps in heavy areas with ENBoost? You see, from my internal point of view after all assumptions tested and failed it's lower performance of basic functions from dx libraries (other wrappers or hooked) or unbeliavable huge cpu cache misses, i have never seen such big performance impact triggered by little bigger functions length. I'm using only fast interlocked thread sync functions and rarely critical sections, memory barrier once per frame cannot affect performance at all. One person posted comparison shots with UsePatchSpeedhackWithoutGraphics true and false toggled, difference is huge (not all users with issue noticed improvement by this parameter, also strange), while code is not that different. In early verions for GTA5 i've used different dx functions and they were extremely slow for AMD (drivers, as always), so i removed almost all Get*** calls from dx and tracking changes myself. But NVidia users have the issue too? I'm completely stuck, without direct access to problematic rig it's like training someone to swim via phone.