Is "LLM" automatically replaced by the "Reflex" if available now? (NVCP)

[Yukai]

I don't currently have Fortnite installed, so I can't check myself, but try Settings > Video > Advanced Graphics.

Apparently Fortnite don't show all graphic settings under "Perfromance" mode, which is based on DX11.
However if u switch to DX11/DX12 rendering mode it'll show everything including Reflex setting. Also if u turn Reflex on and change back to performance -- it still works (based of fps limit).

LLM will only do that if 1) you have Fortnite running in DX11 mode (LLM doesn't work in DX12 or Vulkan), and 2) you're using LLM "Ultra" with G-SYNC + NVCP V-SYNC.

Thanks, didn't know. I checked for myself and I'm sure it wasn't my inattentiveness since on 180hz LLM "Ultra" framerate limit constantly varies around 171-174fps, meanwhile with Reflex, as it was with supposed "glitch", it is dead locked on 171fps. (And I'd never click Ultra since I use in-game 177fps limiter for LLM "On" which I have in NVCP by default (in case games don't run Reflex))

No, LLM is driver-level, Reflex is engine-level.

I thought maybe Reflex just overrites LLM by default if available now, but figures out I just got confused.

So my theory is since it was fresh reinstallment of the game (with saved settings log) it might had some kind of glitch where the game couldn't read my config properly for the first time (or Reflex just wouldn't work for some reason), but after relaunch or two it magically fixed without me turning it on anywhere. Anyway now I know where to change it if needed, guess it's wrapped up! ^-^

[Impecable]

So I've been wondering where the hell Fortnite moved Nvidia Reflex setting since I couldn't find it in game.
I tried messing up with LLM in NVCP and apparently it being "On" limits my fps to 171 (which is standart 180hz Reflex limit)

They are two separate things which causes more latency if both applied.

When a GPU is above 95% usage it goes into a GPU bound state where the frames drawn by the cpu become to much for the GPU to handle, this is when frames get stored into a buffer which adds latency.

Nvidia Ultra Low Latency or NULL aims to reduce that latency when in a GPU bound state, but adds latency when below 95% because NULL has to read the frame buffer which has a latency penalty.

In order to stay below 95% you can use a frame cap to restrict the GPU usage which under game load, Nvidia reflex does a similar thing only the frame cap is dynamic.

Reflex +boost tries to prevent GPU frequency drops in order to maintain low latency but can cause erratic FPS.

So if your using Reflex disable NULL.
A frame cap can replace Reflex but it can take time to find the right FPS target.

Check the fortnite settings/config file and edit it with notepad, you should be able to modify that to enables a setting. (idk where the file is located tho)

NULL will have higher latency than Reflex or a frame cap.

So if I'm getting consistently under 90% GPU usage (fortnite, 4080, 12700k) should I cap my fps to my monitors refresh rate (390) with LLM off and reflex off, or reflex on + unlimited fps cap? And since you said if I'm getting under 95% usage LLM/ULLM should be avoided altogether, correct?

[Tiberiusmoon]

If your FPS can maintain 390FPS in any load conditions in the game then you can disable reflex and LLM, if your FPS dips then enable Reflex because a dip in FPS below 390 will mean there is an increase in GPU usage therefore latency.

Gsync ULLM is use case dependant as it can reduce screen tearing, the latency impact of ULLM is less than 1ms for a 360Hz monitor, I have tested ULLM latency on a 240Hz and 144Hz monitor with 240Hz monitor only increasing latency by 1ms.
https://www.reddit.com/r/apexlegends/co ... x_latency/

The use case in question is the type of game/playstyle your playing, if your playing Val when your hard scoping angles where reaction time is everything latency is key.
BUT if you want to track your enemy to predict movement then visual clarity is key, the reason for this is in certain fast paced scenario's the motion of the game can be distorted by screen tearing which can mess up your predictions.
If your sensitive to motion blur then you could try the BFI options in your monitor.

The higher the monitor refresh rate the low the latency penalty of ULLM and BFI.