Next

G-SYNC 101: Control Panel

Published Jun 20, 2017 by in: G-SYNC

G-SYNC Module

The G-SYNC module is a small chip that replaces the display’s standard internal scaler, and contains enough onboard memory to hold and process a single frame at a time.

The module exploits the vertical blanking interval (the span between the previous and next frame scan) to manipulate the display’s internal timings; performing G2G (gray to gray) overdrive calculations to prevent ghosting, and synchronizing the display’s refresh rate to the GPU’s render rate to eliminate tearing, along with the delayed frame delivery and adjoining stutter caused by traditional syncing methods.

G-SYNC Demo

The below Blur Busters Test UFO motion test pattern uses motion interpolation techniques to simulate the seamless framerate transitions G-SYNC provides within the refresh rate, when directly compared to standalone V-SYNC.

G-SYNC Activation

“Enable for full screen mode” (exclusive fullscreen functionality only) will automatically engage when a supported display is connected to the GPU. If G-SYNC behavior is suspect or non-functioning, untick the “Enable G-SYNC, G-SYNC Compatible” box, apply, re-tick, and apply.

Blur Buster's G-SYNC 101: Control Panel

G-SYNC Windowed Mode

“Enable for windowed and full screen mode” allows G-SYNC support for windowed and borderless windowed mode. This option was introduced in a 2015 driver update, and by manipulating the DWM (Desktop Windows Manager) framebuffer, enables G-SYNC’s VRR (variable refresh rate) to synchronize to the focused window’s render rate; unfocused windows remain at the desktop’s fixed refresh rate until focused on.

G-SYNC only functions on one window at a time, and thus any unfocused window that contains moving content will appear to stutter or slow down, a reason why a variety of non-gaming applications (popular web browsers among them) include predefined Nvidia profiles that disable G-SYNC support.

Note: this setting may require a game or system restart after application; the “G-SYNC Indicator” (Nvidia Control Panel > Display > G-SYNC Indicator) can be enabled to verify it is working as intended.

G-SYNC Preferred Refresh Rate

“Highest available” automatically engages when G-SYNC is enabled, and overrides the in-game refresh rate selector (if present), defaulting to the highest supported refresh rate of the display. This is useful for games that don’t include a selector, and ensures the display’s native refresh rate is utilized.

“Application-controlled” adheres to the desktop’s current refresh rate, or defers control to games that contain a refresh rate selector.

Note: this setting only applies to games being run in exclusive fullscreen mode. For games being run in borderless or windowed mode, the desktop dictates the refresh rate.

G-SYNC & V-SYNC

G-SYNC (GPU Synchronization) works on the same principle as double buffer V-SYNC; buffer A begins to render frame A, and upon completion, scans it to the display. Meanwhile, as buffer A finishes scanning its first frame, buffer B begins to render frame B, and upon completion, scans it to the display, repeat.

The primary difference between G-SYNC and V-SYNC is the method in which rendered frames are synchronized. With V-SYNC, the GPU’s render rate is synchronized to the fixed refresh rate of the display. With G-SYNC, the display’s VRR (variable refresh rate) is synchronized to the GPU’s render rate.

Upon its release, G-SYNC’s ability to fall back on fixed refresh rate V-SYNC behavior when exceeding the maximum refresh rate of the display was built-in and non-optional. A 2015 driver update later exposed the option.

This update led to recurring confusion, creating a misconception that G-SYNC and V-SYNC are entirely separate options. However, with G-SYNC enabled, the “Vertical sync” option in the control panel no longer acts as V-SYNC, and actually dictates whether, one, the G-SYNC module compensates for frametime variances output by the system (which prevents tearing at all times. G-SYNC + V-SYNC “Off” disables this behavior; see G-SYNC 101: Range), and two, whether G-SYNC falls back on fixed refresh rate V-SYNC behavior; if V-SYNC is “On,” G-SYNC will revert to V-SYNC behavior above its range, if V-SYNC is “Off,” G-SYNC will disable above its range, and tearing will begin display wide.

Within its range, G-SYNC is the only syncing method active, no matter the V-SYNC “On” or “Off” setting.

Currently, when G-SYNC is enabled, the control panel’s “Vertical sync” entry is automatically engaged to “Use the 3D application setting,” which defers V-SYNC fallback behavior and frametime compensation control to the in-game V-SYNC option. This can be manually overridden by changing the “Vertical sync” entry in the control panel to “Off,” “On,” or “Fast.”

Next

3700 Comments For “G-SYNC 101”

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Sort by:   newest | oldest | most liked
Roan
Member
Roan

I have a question regarding the recommendation to disable CPU core parking in Windows. Does this still apply in 2025 with the latest version of Win 11 and with a recent Intel Lunar Lake Cpu (Core Ultra 7 258v with Arc 140v iGPU)? The Cpu has 4x performance cores and 4x efficiency cores. I’m using an MSI Claw 8 AI+ handheld PC for gaming. The iGPU outputs to a VRR 48-120hz LFC screen, VRR and vsync enabled globally in Intel graphics software, global framerate cap to 117 with RTSS, as per your recommendations.

I’m getting conflicting info online about CPU core parking on newer Intel chips. Some people are saying that it’s better to leave parking on with the newer chips and latest Windows being designed to and able to work better that way with the P and E cores, whilst some are arguing its better to disable core parking the traditional way for gaming. Some even recommend disabling the E cores in the bios and only use the P cores, but others are against that. Wildly differing recommendations. Keeping in mind this device is basically a small efficient gaming laptop, what would your recommendation be? I guess it comes down to if Windows is able to schedule the correct cores for the correct tasks effectively?

Also, if I wanted to disable core parking, why do I need the ParkControl app when you can disable core parking in the registry manually or by using an app that exposes all the hidden power settings? As far as I understand, it’s set and forget and doesn’t need ParkControl to be installed and launch with Windows every time.

Follow up question – should I keep Cpu scaling at 5-100% like it comes standard on the balanced power setting, or should I push it to full 100% all the time, on AC and battery, for snapper performance but with the added heat and power draw? I would prefer to not push the Cpu to 100% all the time like on my PC, so does Intel Cpu scaling work well nowadays? Negligible performance impact on gaming frametimes/stutter?

My end goal here is to squeeze the best possible performance out of this little machine whilst not needlessly wasting battery life for a miniscule performance bump. But I care more about performance than battery life, when it comes down to it. I play 80% plugged into an outlet.

Like0Vote Down 
2 months 20 days ago
jorimt
Author
jorimt

I’m getting conflicting info online about CPU core parking on newer Intel chips.

Last I commented on core parking was here:
https://forums.blurbusters.com/viewtopic.php?f=10&t=11690&p=96622&hilit=core+parking#p96622

Still using the above config without issue.

Whether it’s optimal for your configuration you’d have to look into specifically, but generally (sans laptop-specific power delivery and battery life concerns), keeping cores unparked is, at best, beneficial, and at worst, neutral.

Also, if I wanted to disable core parking, why do I need the ParkControl app when you can disable core parking in the registry manually or by using an app that exposes all the hidden power settings? As far as I understand, it’s set and forget and doesn’t need ParkControl to be installed and launch with Windows every time.

Ease-of-use and convenience. You’re free to use whatever method you prefer.

Follow up question – should I keep Cpu scaling at 5-100% like it comes standard on the balanced power setting, or should I push it to full 100% all the time

I’m on a dedicated desktop, and I let it scale (5-100%). It will peg itself to its highest base boost clock once a game app is launched anyway, so there should be little to no difference in that scenario.

Like0Vote Down 
2 months 20 days ago
Roan
Member
Roan

One more thing – what is your opinion on setting Intel C-states enabled/disabled in the Bios?

I’ve read that the consensus is generally to keep Intel Speed Shift Tech enabled for best performance and stable frame pacing, but on the topic of C-states people are, again, split on the best approach.

Thanks!

Like0Vote Down 
2 months 20 days ago
jorimt
Author
jorimt

I personally leave them at defaults, but I don’t overclock.

Users that do, tend to adjust those and other settings in the BIOS in attempts to increase overclock stability, typically.

Disabling them in non-overclocked or even mild to moderate overclocked scenarios, especially on a laptop, shouldn’t have a net benefit over on.

Like2Vote Down 
2 months 19 days ago
Roan
Member
Roan

Thank you for the response. Will follow your advice. I had not see the previous discussion about this same topic.

Like1Vote Down 
2 months 20 days ago
Elessar
Member
Elessar

One question I’m hoping you’re able to answer. If I’m playing a game where I can’t hit my refresh rate, and the game doesn’t have a reflex setting (Indiana Jones is my example here), what should I do? Am I reading it correctly that I would do G-Sync + NVCP V-Sync + FPS limit + Low Latency set to On?

That’s what I gather looking at the last bullet point above the power settings section. Just wondering if I’m wrong on that. Thanks

Like1Vote Down 
2 months 23 days ago
jorimt
Author
jorimt

The scenario you’re referring to is GPU-limitation, which increases render queue-level latency separate of the sync latency VRR addresses, so it will occur with or without VRR enabled.

Further, none of my original tests in this article involved GPU-limitation; GPU usage was well below 99% in all scenarios.

In other words, those secondary guidance notes under my Optimal G-SYNC Settings are just there for completion’s sake. Optimal VRR itself is just VRR on + V-SYNC on + exclusive fullscreen-style flip model + framerate at least 3 frames within the refresh rate.

So yes, you can use LLM “On” with VRR in GPU-limited scenarios, but it isn’t as good at reducing render-queue latency as a manual limit set slightly below your lowest achievable average framerate (that keeps GPU usage below 99%) in the given game.

For instance, if I’m playing Indiana Jones at my chosen graphical settings, and my system jumps between 67-110 FPS due to maxed GPU usage in the heavier scenes, and Reflex isn’t available, I’m personally going to leave LLM off and set a manual limit to 60, which will reduce render queue-level latency even more than LLM (assuming said limit keeps GPU usage below 99%) and increase average framerate consistency, but ymmv.

Like0Vote Down 
2 months 23 days ago
Elessar
Member
Elessar

I was just going off the “if framerate does not always reach or exceed refresh rate” which I assume would only ever be in situations of GPU limitation. I think I get what you’re saying though.

So if I don’t want to set a framerate limit to my average, then LLM on is the alternative and should work okay? If I do set a limit to my average fps, is it fine to leave LLM on or would that cause issues?

Like0Vote Down 
2 months 23 days ago
jorimt
Author
jorimt

I was just going off the “if framerate does not always reach or exceed refresh rate” which I assume would only ever be in situations of GPU limitation.

That can happen with CPU limitation as well, but it’s much rarer of a scenario.

So if I don’t want to set a framerate limit to my average, then LLM on is the alternative and should work okay?

Reflex > manual FPS limit > LLM for GPU-bound scenarios.

LLM is passable in that instance, but unlike Reflex or a manual limit, which entirely prevent extra pre-rendered frames, LLM will only reduce them, mitigating render queue latency by 1 to 1 1/2 frame whenever the GPU usage is maxed, typically.

If I do set a limit to my average fps, is it fine to leave LLM on or would that cause issues?

LLM tries to forcibly starve the render queue, so in the worst of cases, it can trigger longer stutter during framerate spikes, especially when the game isn’t GPU-limited.

Like0Vote Down 
2 months 23 days ago
Elessar
Member
Elessar

Thanks for answering. Looking for a clarification on the last point, if I do decide to go the route of limiting to average fps, I should turn LLM off because it could cause issues?

For Indiana Jones I was previously using an fps cap just below refresh rate (this did nothing I guess since I don’t reach it), then V-Sync and G-Sync on. This is fine for eliminating screen tearing right? But Reflex, or a framerate limit at average FPS, or LLM would all lower the input delay which is the main thing these things aim to lower?

Lastly (sorry, I have a habit of asking too many questions, just trying to get this down), if I do want to go LLM, then should I keep a framerate cap just below monitor refresh rate?

Like0Vote Down 
2 months 22 days ago
jorimt
Author
jorimt

It’s important to understand what each of the settings you’re using are actually doing, and then you’ll know how and when to mix-and-match them situationally.

The term “LLM” is present 800+ times in this comments section, so for future reference, it can be easily page searched after clicking “Load All Comments” at the bottom of this section.

Versions of what I’m about to re-explain are already available in my article, as well as multiple times in the comments section, so wall-of-text warning…

G-SYNC + V-SYNC:
Prevents tearing and sync latency whenever the framerate is (typically, at least 3 frames) within the currently set physical refresh rate.

VRR does not prevent or reduce render queue latency whenever the GPU usage is 99%+, nor will the minimum -3 FPS limit (intended only to keep the framerate within the refresh rate to ensure VRR remains engaged in the cases the framerate could otherwise exceed the refresh rate) if the GPU usage reaches or exceeds 99% at said limit.

Reflex:
Prevents render queue latency (with or without VRR, whether the framerate is above or below the refresh rate) caused from 99%+ GPU usage by dynamically monitoring said usage, and then setting an automatic limit slightly below the achievable average framerate.

Reflex does not directly prevent sync latency, and its only relation to G-SYNC is the automatic limit it sets slightly below the refresh rate when combined with G-SYNC + NVCP V-SYNC to ensure VRR remains engaged (which is a substitute to my minimum manual -3 FPS recommendation; LLM “Ultra” does this too).

When G-SYNC + V-SYNC + Reflex is enabled and the average framerate naturally drops below the auto limit due to more demanding scenes causing 99%+ GPU usage, Reflex switches to the dynamic limiting behavior I described earlier above.

Manual FPS limit:
(that keeps GPU usage below 99% at all times)
Same effect as Reflex’s dynamic non-VRR limiting mechanism, just with a statically set limit.

LLM “On:”
Unlike Reflex, in supported games/engines, if the max pre-rendered frames aren’t already internally set to “1,” LLM will shrink the max pre-rendered frames from the default driver-level setting of “3” to “1,” meaning the CPU can only prepare up to 1 instead of 3 frames in advance whenever the GPU usage is 99%+, which will typically reduce render queue latency by 1 to 1 and 1/2 frames over LLM being off in GPU-limited scenarios (it won’t technically do anything in non-GPU-limited scenarios), but since LLM is starving the render queue, if your particular CPU can’t keep up as well with less “backup” frames in the render queue while the GPU is busy, a previous frame will repeat delivery instead, causing the appearance of stutter.

LLM “Ultra” is similar to “On,” except instead of reducing the max pre-rendered frames to “1,” it tries to make that an effective “0” by delivering the most recently CPU-prepared frame “just-in-time” whenever the GPU usage is 99%+, but if it can’t (and it often doesn’t), it will repeat a previous frame instead, causing stutter.

This is why I say “LLM” is inferior to Reflex or a manual limit for reducing render queue latency, since Reflex/a manual limit will entirely prevent the GPU usage from maxing in the first place, while LLM will only attempt to mitigate the issue after GPU usage is already maxed.

LLM is essentially a very weak and ineffective setting with limited use-cases, and may negatively impact frametime performance vs. off more often than not.

I don’t personally recommended it in any use-case, but it’s up to the individual to decide for themselves.

Like0Vote Down 
2 months 22 days ago
Elessar
Member
Elessar

Thanks, I think I understood most of these but it is very helpful for you to lay it out, I was just wondering what the hierarchy of best to worst would be in a game where you aren’t hitting you refresh rate (so presumably GPU bound) vs one where you are to get the best experience factoring in tearing, latency and stuttering (most important imo)

So if above refresh rate would that look like:

G-Sync, V-Sync, FPS limit -> G-Sync, V-Sync, Reflex -> Anything else

And below refresh rate:

G-Sync, V-Sync, Reflex -> G-Sync, V-Sync, manual limit set to average -> G-Sync, V-Sync, LLM On -> G-Sync, V-Sync?

Would the potential stuttering from LLM be worse than just running without?

Like0Vote Down 
2 months 22 days ago
jorimt
Author
jorimt

I was just wondering what the hierarchy of best to worst would be in a game where you aren’t hitting you refresh rate (so presumably GPU bound) vs one where you are to get the best experience factoring in tearing, latency and stuttering (most important imo)

I already answered you in an earlier reply with:

Reflex > manual FPS limit > LLM for GPU-bound scenarios.

LLM is the last/worst option for GPU-limited scenarios, and needn’t be used for non-GPU-bound scenarios.

Now that you have the information on what each does and how they do it, it’s up to you where, when and if to use them.

Would the potential stuttering from LLM be worse than just running without?

Whether any LLM-caused stutter is a deal-breaker for you is only for you to decide; try and see.

Just a fair warning; there’s no reliable way to validate that the pre-rendered frames reduction component of LLM is even engaged in a given game; some games it works in, others it does nothing when enabled, so you kind of just have to activate it and hope it does something.

Like1Vote Down 
2 months 21 days ago
user2422
Member
user2422

may i ask what your personal settings are regarding llm, reflex, in-game vs external fps cap for both cpu and gpu bound scenarios ?

Like0Vote Down 
2 months 25 days ago
jorimt
Author
jorimt

I leave LLM off.

I typically only use Reflex whenever my system can’t reach my target average framerate 99% of the time at my chosen graphical settings, so it will 1) keep the framerate within the refresh rate via its auto limit when combine with G-SYNC on + NVCP V-SYNC, and 2) prevent render queue latency due to maxed GPU usage whenever the framerate drops below it.

I.E. I’ll sometimes sacrifice a locked x FPS for a slightly fluctuating x FPS with higher graphical settings and let Reflex take care of the rest.

If, however, my system can reach my target average framerate + graphical settings 99%+ of the time without maxing GPU usage, I disable Reflex and use an appropriate in-game or external limiter instead.

As to whether I use an in-game or external limiter, and which external limiter I use, it’s entirely dependent on the game.

Like1Vote Down 
2 months 24 days ago
user2422
Member
user2422

do you make any further changes in the nvidia profile inspector? like setting max pre rendered frames to 1 (if not using llm) or anything like that?

and in those gpu bound scenarios you mentioned do you use reflex on or on + boost? (i know the differences, just curious what your setup is with all your years worth of knowledge)

and when reflex is not available in that gpu heavy game do you just lower the graphics/framerate limit until you prevent 99% utilization?

Like0Vote Down 
2 months 24 days ago
jorimt
Author
jorimt

do you make any further changes in the nvidia profile inspector? like setting max pre rendered frames to 1 (if not using llm) or anything like that?

I typically apply V-SYNC (when using G-SYNC) and Prefer maximum performance per game profile via NVI (to ensure my GPU is fully utilized in less performance intensive games), and the Nvidia MFR limiter if I’m not using another.

I’ll also occasionally apply reBAR in games that don’t have it enabled at default that are also community verified to actually improve performance (rare).

And that’s about it, unless I’m playing a 20+ year old game that I want to apply some archaic AA (SGSSA etc) or HBAO+ to, or the odd game that only supports legacy borderless windowed mode in VRR.

and in those gpu bound scenarios you mentioned do you use reflex on or on + boost? (i know the differences, just curious what your setup is with all your years worth of knowledge)

I rarely use Reflex, and try instead to use a manual average framerate limit that my system can sustain 99% of the time.

I.E. I prefer a locked (VRR) framerate over one that fluctuates; if my system can get 120-215 FPS in a more demanding SP game, I’d rather lock it to 120 FPS than let it drift all over, or with 60-72 FPS, I’d rather a solid 60 FPS, etc, but this is highly preferential, not strictly recommended.

and when reflex is not available in that gpu heavy game do you just lower the graphics/framerate limit until you prevent 99% utilization?

Pretty much.

Like1Vote Down 
2 months 24 days ago
user2422
Member
user2422

so you only use vsync per game with inspector and not globally? if so, why?

Like0Vote Down 
2 months 24 days ago
jorimt
Author
jorimt

Read this comment and mine below it:
https://forums.blurbusters.com/viewtopic.php?f=5&t=14525&p=115608&hilit=global+v+sync#p115486

It’s just to rule out any modern weirdness (MPO, etc). Not something easy to blanket recommend, as it can be situational, and may even change from OS/driver updates over time.

Like2Vote Down 
2 months 24 days ago
user2422
Member
user2422

thanks for the answers.

Like2Vote Down 
2 months 24 days ago
mike-lesnik
Member
mike-lesnik

Hi jorimt! Thanks for your answers!
What impact does LFC have on inputlag (with all your recommendations)? For example, 57 frames VS 56+LFC on 165-240Hz g-sync compatible monitor.
Can LFC make stuttering worse?
Is there any difference between LFC on LCD and OLED?

Like0Vote Down 
2 months 26 days ago
jorimt
Author
jorimt

What impact does LFC have on inputlag (with all your recommendations)? For example, 57 frames VS 56+LFC on 165-240Hz g-sync compatible monitor.

LFC shouldn’t affect latency (your example scenario included).

Can LFC make stuttering worse?

Depends on whether it’s G-SYNC native or G-SYNC Compatible FreeSync, and where the LFC threshold starts.

Is there any difference between LFC on LCD and OLED?

Not functionally, but again, where LFC starts is typically model + refresh rate-dependent.

Like0Vote Down 
2 months 25 days ago
mike-lesnik
Member
mike-lesnik

Depends on whether it’s G-SYNC native or G-SYNC Compatible FreeSync, and where the LFC threshold starts.

Will there be a noticeable difference in frametime stability if we compare “VRR off” vs “G-Sync compatible” with LFS threshold 58Hz? In conditions where fps regularly drops below 50.

Like0Vote Down 
2 months 25 days ago
jorimt
Author
jorimt

With a monitor that has proper LFC behavior, no, there shouldn’t be.

Like0Vote Down 
2 months 24 days ago
tummySkaren
Member
tummySkaren

Hi, I wanted to double check that the settings here (https://blurbusters.com/gsync/gsync101-input-lag-tests-and-settings/14/) are optimal for my set-up given the following:
1. I only play eSports games like Valorant, Overwatch 2, where I get 400fps+ at least at all times, much higher than my 240hz refresh rate.
2. I’m using an OLED monitor that is G-Sync Compatible (AW2725Q)
3. I don’t know whether absolute lowest latency or absolute highest consistency is better, I believe motion clarity is the most important? Not really sure

Like1Vote Down 
2 months 29 days ago
jorimt
Author
jorimt

My Optimal G-SYNC settings ensure the lowest possible tear-free latency for framerates within the refresh rate, so if that’s your goal, use them.

Whether they’re optimal for OLED and/or comp play at ultra high framerates is up to the individual user to decide.

OLED is prone to VRR flicker, for instance, and if you’re running uncapped 400+ FPS at 240Hz, properly configured G-SYNC will not accommodate that scenario, since VRR only functions within the refresh rate.

In other words, if you’re primarily playing uncapped comp games on an OLED display, G-SYNC off + V-SYNC off + maximum framerate that does not max GPU usage is the lowest possible latency scenario, and thus probably best for your use case. Entirely up to you though.

Like0Vote Down 
2 months 26 days ago
wpDiscuz

Related Content

Featured Content

Blur Buster Forums

Blur Busters Forums

TestUFO Motion Tests

TestUFO Motion Tests