G-sync refresh rate/fps question.(long read)

[hamsterman]

I was checking my gpu and cpu usage and it's completely normal but my fps would drop from 158 to around 100 and then it happens briefly.

If that results in the game looking like it's momentarily visually pausing, that's frametime spikes.

And in that case, It's not only dropping to, say, 100 FPS, it's actually dropping to 0, but system-side framerate counters typically only report every second, so they're behind enough not to reflect the literal drop.

And, no, you don't want the counter to update fully in real-time, or it will cause further performance issues due to pinging too frequently, interrupting the game and other processes.

You instead want to look at the frametime graph; if there's an [x]ms spike, that means the same frame was repeated (aka stutter) for [x]ms until the next new frame was displayed.

If however, what you're experiencing feels more like continual jitter, that's frametime variance (sometimes more commonly referred to as "micro-stutter"), which can be mitigated by ensuring you have a framerate limit set that your system can consistently reach 99% of the time. And if that doesn't fix it, it means the game itself has internal frame pacing issues that can't easily be corrected on the user-side (or at all in some cases).

I've been doing some tests. I've noticed with vsync enabled and gsync enabled with reflex + boost when I'm in a game during intense parts my fps drops, I notice incredible input delay. Such as swinging my weapon it feels very sluggish and slow. Is this because of vsync? I tried disabling vsync and played the game again and during the fps drops I feel like the delay was improved. I don't know if this is placebo or what is causing this.

Also inside the game there is a DLSS option, I should always have this enabled right? What about super resolution too? https://imgur.com/a/HVRkEOk

[jorimt]

I've been doing some tests. I've noticed with vsync enabled and gsync enabled with reflex + boost when I'm in a game during intense parts my fps drops, I notice incredible input delay. Such as swinging my weapon it feels very sluggish and slow. Is this because of vsync?

I tried disabling vsync and played the game again and during the fps drops I feel like the delay was improved. I don't know if this is placebo or what is causing this.

The only functional difference between G-SYNC on + V-SYNC on and G-SYNC on + V-SYNC off, is the former can't tear and the latter can. So whenever you see a tear with the latter, that's your one and only literal input lag reduction; the tear itself.

To be clear, lower frame/refresh rates feel inherently more sluggish since there are less updates per second. G-SYNC can't fix this, and while Reflex can reduce the extra latency in scenarios where the framerate is dropping due to maxed GPU usage, it cannot remove the feeling of sluggishness caused by the lower average framerate itself.

Also inside the game there is a DLSS option, I should always have this enabled right? What about super resolution too? https://imgur.com/a/HVRkEOk

Base DLSS is an upscaling technique not directly related to G-SYNC, V-SYNC, or Reflex (and Reflex isn't directly related to G-SYNC or V-SYNC either). Whether you have it enabled is up to you. DLSS reduces the internal game resolution, and then upscales that resolution with advanced techniques to make it appear more like the resolution it is upscaling to.

DLSS will reduce the load on the GPU, making the game more CPU-reliant.

As for "Super Resolution," that's part of DLSS4, separate of upscaling, which, according to Nvidia:
https://developer.nvidia.com/rtx/dlss?s ... itle%3Aasc

boosts performance by using AI to output higher-resolution frames from a lower-resolution input. DLSS samples multiple lower-resolution images and uses motion data and feedback from prior frames to construct high-quality images.

[hamsterman]

I've been doing some tests. I've noticed with vsync enabled and gsync enabled with reflex + boost when I'm in a game during intense parts my fps drops, I notice incredible input delay. Such as swinging my weapon it feels very sluggish and slow. Is this because of vsync?

I tried disabling vsync and played the game again and during the fps drops I feel like the delay was improved. I don't know if this is placebo or what is causing this.

The only functional difference between G-SYNC on + V-SYNC on and G-SYNC on + V-SYNC off, is the former can't tear and the latter can. So whenever you see a tear with the latter, that's your one and only literal input lag reduction; the tear itself.

To be clear, lower frame/refresh rates feel inherently more sluggish since there are less updates per second. G-SYNC can't fix this, and while Reflex can reduce the extra latency in scenarios where the framerate is dropping due to maxed GPU usage, it cannot remove the feeling of sluggishness caused by the lower average framerate itself.

Also inside the game there is a DLSS option, I should always have this enabled right? What about super resolution too? https://imgur.com/a/HVRkEOk

Base DLSS is an upscaling technique not directly related to G-SYNC, V-SYNC, or Reflex (and Reflex isn't directly related to G-SYNC or V-SYNC either). Whether you have it enabled is up to you. DLSS reduces the internal game resolution, and then upscales that resolution with advanced techniques to make it appear more like the resolution it is upscaling to.

DLSS will reduce the load on the GPU, making the game more CPU-reliant.

As for "Super Resolution," that's part of DLSS4, separate of upscaling, which, according to Nvidia:
https://developer.nvidia.com/rtx/dlss?s ... itle%3Aasc

boosts performance by using AI to output higher-resolution frames from a lower-resolution input. DLSS samples multiple lower-resolution images and uses motion data and feedback from prior frames to construct high-quality images.

I've noticed for this specific game if I cap my FPS within the nvcp app to 144 the game feels somewhat better. In this situation would vsync be on or off? Sorry if I repeated myself it's a lot of information and I'm slowly starting to understand. I tried vsync off and I don't believe I get any tear at 144 cap, I will test again.

As far as I understand vsync should always be on with gsync no matter the cap? I mange to get 158 fps no problem in this game but in certain intense situations my fps can drop as low as 90-100. Changing and capping to 144 just makes it better, or am I misunderstanding.

Thank you for the help.

[jorimt]

As far as I understand vsync should always be on with gsync no matter the cap?

G-SYNC on + V-SYNC on vs. G-SYNC on + V-SYNC off is my most frequently asked question, and is answered in more detail in the second entry of my Closing FAQ:
https://blurbusters.com/gsync/gsync101- ... ttings/15/

Simplest explanation is, G-SYNC on + V-SYNC on IS G-SYNC, whereas G-SYNC on + V-SYNC off is still G-SYNC, but with a tearing-prevention component removed.

G-SYNC dynamically steers the tearline off-screen based on the current framerate for framerates within the refresh rate.

G-SYNC on + V-SYNC on does this regardless of frametime performance, whereas G-SYNC on + V-SYNC off only does this when the frametime performance is stable enough; whenever it’s not, G-SYNC reverts to no sync and tears instead of syncing (even when the framerate is within the refresh rate).

So whenever you see a tearline with G-SYNC on + V-SYNC off, that’s the only latency “reduction” and/or smoothness difference you’ll be getting over G-SYNC on + V-SYNC on. Thus, any “latency” with G-SYNC on + V-SYNC on IS the lack of tearing, whereas the latency “reduction” with G-SYNC on + V-SYNC off is the TEARING itself.

I mange to get 158 fps no problem in this game but in certain intense situations my fps can drop as low as 90-100. Changing and capping to 144 just makes it better, or am I misunderstanding.

Manually capping your framerate to a limit your system can sustain an average of 99% of the time increases the feeling of consistency because it prevents wild swings in framtime/frame/refresh rate.

So you're free to set the limit to whatever you wish within the refresh rate when using G-SYNC + V-SYNC, and then G-SYNC will ensure that framerate is tear-free without adding it's own sync-related stutter or latency on top of that.

But to be clear, games can and often will still stutter and/or lag with or without G-SYNC.

[hamsterman]

As far as I understand vsync should always be on with gsync no matter the cap?

G-SYNC on + V-SYNC on vs. G-SYNC on + V-SYNC off is my most frequently asked question, and is answered in more detail in the second entry of my Closing FAQ:
https://blurbusters.com/gsync/gsync101- ... ttings/15/

Simplest explanation is, G-SYNC on + V-SYNC on IS G-SYNC, whereas G-SYNC on + V-SYNC off is still G-SYNC, but with a tearing-prevention component removed.

G-SYNC dynamically steers the tearline off-screen based on the current framerate for framerates within the refresh rate.

G-SYNC on + V-SYNC on does this regardless of frametime performance, whereas G-SYNC on + V-SYNC off only does this when the frametime performance is stable enough; whenever it’s not, G-SYNC reverts to no sync and tears instead of syncing (even when the framerate is within the refresh rate).

So whenever you see a tearline with G-SYNC on + V-SYNC off, that’s the only latency “reduction” and/or smoothness difference you’ll be getting over G-SYNC on + V-SYNC on. Thus, any “latency” with G-SYNC on + V-SYNC on IS the lack of tearing, whereas the latency “reduction” with G-SYNC on + V-SYNC off is the TEARING itself.

I mange to get 158 fps no problem in this game but in certain intense situations my fps can drop as low as 90-100. Changing and capping to 144 just makes it better, or am I misunderstanding.

Manually capping your framerate to a limit your system can sustain an average of 99% of the time increases the feeling of consistency because it prevents wild swings in framtime/frame/refresh rate.

So you're free to set the limit to whatever you wish within the refresh rate when using G-SYNC + V-SYNC, and then G-SYNC will ensure that framerate is tear-free without adding it's own sync-related stutter or latency on top of that.

But to be clear, games can and often will still stutter and/or lag with or without G-SYNC.

Thank you. I've started to understand now. I will cap my fps at 144 within this specific game and keep vsync on in nvcp app. I will also use ultra low latency mode if the game does not support it, thank you!

[wiseude]

Apparently the only game out there in recent years to not have spikes and stutters is KCD2 These days, saying that you're getting stutters and frametime spikes in games is like saying the sky is blue.

https://www.youtube.com/watch?v=w31Ct25gJlI

Out of curiosity.Is it normal for the refresh rate to frequently hit 140+ when I'm playing a game with an fps limiter of 120?
I noticed when playing BG3 and enabling the refresh rate overlay on my monitor (Dell S2721DGFA) it stays around the 120 ballparks but every couple of seconds or when I move the camera around I notice it hit 135-144hz for a split second and go back down.How is it possible when my ingame refresh rate is set to 120?

VRR matching the 120 fps and staying in the that ballpark (with an fps limit of 120) makes sense.
Frequently seeing the refresh rate going as high as 135+ when there's an fps limit of 120 doesn't make sense because my fps maximum should be 120 because of the limiter.

Is this intended or is there something going on?

[jorimt]

I noticed when playing BG3 and enabling the refresh rate overlay on my monitor (Dell S2721DGFA) it stays around the 120 ballparks but every couple of seconds or when I move the camera around I notice it hit 135-144hz for a split second and go back down.How is it possible when my ingame refresh rate is set to 120?

It means your monitor-side refresh rate meter is momentarily resetting itself during a large/abrupt enough frametime variance or spike. I.E. its readout can essentially lag behind and must periodically catch up.

It isn't reliable on a moment-to-moment-level, especially on G-SYNC Compatible FreeSync displays.

Ignore it, and reference the system-side readouts instead.

[wiseude]

I noticed when playing BG3 and enabling the refresh rate overlay on my monitor (Dell S2721DGFA) it stays around the 120 ballparks but every couple of seconds or when I move the camera around I notice it hit 135-144hz for a split second and go back down.How is it possible when my ingame refresh rate is set to 120?

It means your monitor-side refresh rate meter is momentarily resetting itself during a large/abrupt enough frametime variance or spike. I.E.

Oh.Wouldn't that show up as a frame drop or micro stutters in a frametime graph then?I see it spike to 135+hz pretty consistently even staying still with a pretty flat frametime graph.

[RealNC]

Oh.Wouldn't that show up as a frame drop or micro stutters in a frametime graph then?I see it spike to 135+hz pretty consistently even staying still with a pretty flat frametime graph.

The only frame limiter I know of that's able to keep the OSD Hz indicator stable is the SpecialK "normal" limiter. Anything else, including RTSS, has fluctuations. This also requires that the game isn't stuttering.

[jorimt]

Oh.Wouldn't that show up as a frame drop or micro stutters in a frametime graph then?

Not necessarily. My general point is these monitor meters are volatile and easily thrown off.

They're essentially toys, and as I've shared multiple times in this forum and elsewhere, the only thing they're good for is determining whether VRR is currently engaged; fluctuating = engaged, static at refresh rate number = disengaged.

They should not be used to validate anything else, including frametime performance or moment-to-moment average variable refresh rate.

I experienced monitor meter spikes similar to what you're reporting all the time on a native G-SYNC display during my original G-SYNC 101 tests, and they never reflected in the actual latency readings.