Do you use Nvidia's Driver Limiter or RTSS to limit your framerate? Any pros and cons to either?

RTSS "hacks" into the rendering pipeline to achieve Frame Limiting, while the built-in Nvidia Framecap does it natively. There's even a warning in the app itself how some anti cheat engines diagnose that as cheating. I fail to understand how RTSS could be superior, unless the Nvidia implementation is done very poorly? [There is some question of why they needed v1, v2, v3 to get it right?]

I have a Gsync monitor too, Vsync globally forced ON. That's also a kind of frame limiting. Is it the same as the v3 frame limiter? I don't think so, because vsync is much older? How does THAT limit frames? Heh

 

I prefer RTSS because I use its OSD anyway with Afterburner for undervolting and custom fan curve. These features are not present in NVCP and I doubt they ever will be - at least not in the current version of NVCP.

As for the NV limiter I did some tests lately and it is now working OK even in OGL and VULKAN. Doesn't work in DX8 and olders but thats not a big deal. One weird thing is that Power management mode must be set to Prefer maximum performance, with Optimal or Adaptive the GPU will not boost to maximum frequency which can lead to stuttering and frame drops. At first I thought its some sort of bug in NV limiter but it seems to be a feature as described here:
https://www.nvidia.com/en-gb/geforce/news/nvidia-geforce-ces-2020-game-ready-driver/

When using RTSS limiter the GPU uses always full range of frequencies even in Optimal mode, just like without any limiter.

 

Weird, my 3080 will boost to max even when set to optimal and using the NV limiter.

EDIT - To clarify, it will boost to the max clock if the game demands the performance within the scope of the frame limit. It's dependent on how close your max clock is to the frame limit you have set. Also, I run my 5700X on balanced, always have and it performs perfectly. No need to change plan.

 

It's weird behavior. Strange Brigade D3D12 clocks up way less aggressively than Vulkan when both used with driver limiter (at least last time I checked).
So ideally always set to maximum performance per profile (and maximum performance CPU power plan can be good idea too with fps limiters).

 

I am using the nVidia CP's limiter ever since it became available and got tested by some reputable sources (to have similar latency penalty compared to RTSS) because I have had issues with RTSS lately (think in a few year's span).

 

I just use a global fps cap of 162 in Rivatuner.
Any games that don't support that, Destiny 2 for example, I cap in game.

Cap 3fps under limit for gsync so its always enabled.
You can use Nvidia low latency stuff too but that will cap your gpu under max usage.

 

In many game's it's almost for free.

 

I’ve not heard before the recommendation to force the max power plan for the CPU in conjunction with FPS limiters — if I recall right for Ryzen people always said to make sure your chipset drivers were up to date then they recommended using the “balanced” option, but I don’t have a clue what the best approach really is since I don’t understand the underlying differences and I’ve not actually tested that approach.

Does the max perf power plan let the CPU run better when capping framerates then? Perhaps I’ll test it out in some CPU heavy games like Spider-Man. Thanks,

 

Why would you not want max performance of from cpu in your games providing the game can use all cores/threads? using balance and power saver and what not have algorithms that dictate if the power is needed, which imo more time then not cause problems. which same reason why i have max performance set globally for my gpu. And before some says it my gpu drops to idle speeds and temps just fine when system is idle, anyone that say max performance keeps the gpu from idle speeds and temps, need to learn how to setup profiles on there system.

 

... but you are using a frame limiter? Do you not see the contradiction?

If you are not using a frame limiter and don't understand why we are - my card burns twice as much power as yours does under full load. Full load is usually not necessary to achieve my 117 FPS target.
Rendering more than 117 FPS which my display can't even show anyway is a waste of power and creates additional heat too, which has to be cooled away.

As for CPU power plan, personally i run the default balanced and i've never (in decades of CPUs) experienced any kind of CPU clocking stutter. But perhaps people running 200+ FPS screens perceive them? We are talking about a kind of people who have frame time overlays open when playing [benchmarks].

 

"We" use fps limiters to stay within VRR range with lower input lag than vsync. To use it for other purposes is just an ugly hack (ok, opinionated, but still).

 

I use it in conjunction with Gsync to get stable/“perfect” frametimes as some games don’t seem to run smoothly uncapped for whatever reason. It really depends though, other games seem to not work well when capped externally and produce weird camera jitter. If only every game could be like DOOM 2016/Eternal or Red Dead 2/Cyberpunk — those games all worked flawlessly/very smoothly regardless (capped, uncapped, whatever).

 

In case of HZD, it also happens with in-game limiter.

 

I already stated i dont, less there pacing issues also Emulators dont dont like adaptive gpu setting anymore then they dont like balanced for cpu. i also dont care to have issue pop up cause of either of the 2, which still happen. and there pc game dont dont work right with adaptive.

It easier to set game globally to max peformance and modifey the few profiles/exe that need to be set to adaptive for idle clocks to work then vis versa, at lest for me anyway

and record i run everything at 60hz60fps even when had the 120+ hz monitor i limited it to 60fps

Limiter in my case is used ONLY when i have pacing issues, which get fixed by it, and it will be used to limit a 120hz or higher monitor to 60 fps, if ever buy such monitor again, the extra heat and extra money need to power such framerates are not worth it to me

 

What I have seen from RTSS is that it's way more precise that the nvidia limiter, even if it's also very good now, I tested this in some games with the frametime graph in real time while playing, RTSS has no odd spikes unless there is some "event" from the game such as a small cutscene, opening the menu, etc, whereas the nvidia limiter had odd, small frametime spikes for no apparent reason, even on the main menu in some games, but RivaTuner super high precision limiter can potentially cause bigger stutters in certain games that can't handle such a strict framerate limit, so in this case the one from nvidia can yield better results, but in general I prefer RTSS.

 

it situational, depending on game, system hardware, drivers and config in my experience

For me in Swtor the ingame "limiter" never worked, RTSS worked better then Nvidia Limiter to smooth out studders atlest when engine wasnt the cause. and with witcher that was studdery mess for me, and i had to force vsync it didnt have ingame vsync and 1 pre-render. and Nvidia limiter to smooth out, where RTSS smoothed it out but not as good as NVidia limiter. But that is just with my system and config mileage will vary It situational in my experience. I tend to use Nvidia limiter more then AB RTSS simply cause i dont like relying on 3rd party programs for something that can be done in drivers. will use in situations calls for it none the less and that just experience with current system, with my previous what worked with that i7 920 system didnt work same with current 6700k

 

It depends on the game engine and overall resources usage I guess. Don't know if things are different with 3xxx GPUs, but that is what I observed with my 2080 and NV limiter with Optimal power - sometimes it boosted, but often it was too late with nasty frame drops. Anyway I use desktop PC for games at 141 FPS (VRR) so I couldn't care less about some notebook's Optimal setting, lol. As for the CPU power plan I have no reason to use Balanced for gaming, my i7 has exactly the same temps as with Max. power and again with Balanced and some retro games bound to one core I get stutters and framedrops. Over the years I become alergic to stutter of any kind, especially with high refresh monitor, so I use my optimal settings for games, which is max. performance + RTSS limiter.

tsunami231 you are really missing out when limiting 120Hz monitor to 60FPS. I could not go back to 60Hz even if I try, feels like slideshow.

 

I was using RTSS (60 FPS) until recently until i found out that caused a lot of retro games to stutter every 3 seconds. Disabling the limiter fixed it, so i decided to just uninstall RTSS all together.

I always have V-Sync (60 Hz) enabled which is my limiter then anyways.

Seems limiting FPS with an external tool does more harm than good (not using the CPL limiter either).

 

I'm curious, which games were affected? I tend to play a lot of games from my childhood, so I could try to reproduce it.

 

TheDeeGee: I doubt it had anything to do with RTSS limiter - maybe some demanding monitoring sensor you activated in OSD? I have a lot of old 3D games and never encountered anything like that (Far Cry 1, FEAR, Max Payne, Quake 1/2, GZDoom, Doom 3, Quake 4, Half Life 2, Unreal 1, UT2004 etc). In fact it was always the opposite - perfect frametimes and zero stutter with RTSS limiter but stuttering/hitching without it. For me some kind of limiter is mandatory to stay in the VRR range, VSync latency sucks as often do inacurrate in-game timers and limiters.