Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Talk about NVIDIA G-SYNC, a variable refresh rate (VRR) technology. G-SYNC eliminates stutters, tearing, and reduces input lag. List of G-SYNC Monitors.
wtfisgoingon
Posts: 6
Joined: 08 Nov 2021, 11:11

Re: Monitor reported refresh rate vs system

Post by wtfisgoingon » 12 Nov 2021, 16:50

RealNC wrote:
09 Nov 2021, 12:20
wtfisgoingon wrote:
08 Nov 2021, 11:59
I have more details I can provide if anyone wants them but right now after not being able to use G-Sync I really just want to spend all day playing with it :)
Thanks for taking the time to post this. You might want to report this to nvidia:

https://nvidia.custhelp.com/app/answers ... -to-nvidia
Thanks I will do that!

wtfisgoingon
Posts: 6
Joined: 08 Nov 2021, 11:11

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by wtfisgoingon » 12 Nov 2021, 16:54

Another couple of updates:

1) I can untick then retick the box while the game is running and it will still fix it. I'm not sure if this is a new thing or because of the last Win 10 update from a couple of days ago because I hadn't ever bothered to try doing it this way. All the other times I would shut down the game, untick appy, retick apply, then start game.

2) This exact behavior happens on my G-Sync compatible monitor (HP 25x) when I allow the monitor to be put into sleep mode and then try to use it afterwards. I had noticed that using it after it being in sleep mode would make it "feel" weird so a couple of years ago I got in the habit of always shutting the monitor off manually to avoid this. I wondered if this was perhaps the same thing happening so I tested allowing it to go into sleep mode and then doing the untick apply retick apply to see if it would correct it and it does.

User avatar
Chief Blur Buster
Site Admin
Posts: 11653
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by Chief Blur Buster » 16 Nov 2021, 15:01

I wonder if there's an API that can do the equivalent of untick/reticking the box.
As a workaround for these types of driver bugs.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

wtfisgoingon
Posts: 6
Joined: 08 Nov 2021, 11:11

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by wtfisgoingon » 16 Nov 2021, 16:47

Chief Blur Buster wrote:
16 Nov 2021, 15:01
I wonder if there's an API that can do the equivalent of untick/reticking the box.
As a workaround for these types of driver bugs.
I wasn't 100% sure if I should put this here or make a new post..

I'm making a video to submit to Nvidia as well as to use to help out others with the same issue.

I'm wondering if anyone interested would mind watching it and letting me know if there is anything else I could add or change or do to improve the video such as other or different RTSS / Afterburner settings or any other program I could run at the same time to show better stats and things like that.

In the video I start recording after a full restart and then load up the game I play the most (although all of these exact same things happen on the other 5 or so games I've tried it on) and let it sit for a few seconds after getting to main menu then I change to different menus because they always create an FPS drop.

Then I alt-tab and shut down RTSS, wait a couple of seconds then restart it and alt-tab back to the game to show that nothing changes other than "Sync Timeout" which will change on each startup of RTSS & while things are "Wonky" this number will always change and varies wildly between ~6.930ms - 24.756ms When things are working normally this number will still change but only very slightly as it will always be within ~6.930ms - 6.960ms. These changes also line up with what the OSD & "info" button on my monitor reports.

Then again after switching between a few menus and pausing a few seconds in between I alt-tab and check & then uncheck the box then alt-tab back into the game and switch between menus & alt-tab out to shut down RTSS then restart it etc to show the differences.

Here is a link to the video: https://youtu.be/MHiHR71DxI4

Some notes:

Every setting is identical the only thing altered is unticking the box, clicking apply, reticking the box, clicking apply. I get the exact same results if I shut down the game then do those steps and then restart the game as I do if I leave the game running so to keep the video short I just left the game running.

If I restart my system things are always "Wonky" until I untick the box apply it and retick the box however after doing this provided that I do not restart my system I can stop & play the same game or other games and things work normally. Again, this happens 100% of the time.. on system startup things are always "Wonky" and after unticking/reticking things will work normally regardless of how many hours pass or how many games I play, movies I watch etc until I restart and then it's "Wonky" again.

I know the OSD & the "info" button on a monitor isn't exactly accurate but I thought it was worth mentioning that when things are wonky the OSD fluctuates wildly between 40-144Hz so quickly it isn't readable & the "info" button will always report a different refresh rate between 40-144Hz each time it's used. When things are normal the OSD always displays a solid 144Hz & the "info" button always reports 144Hz. This happens 100% of the time, regardless of my FPS, it doesn't matter if it is steady or constantly fluctuating.

Using FSE vs Borderless Fullscreen doesn't change anything other than "Present Latency" is higher at all times while using FSE.
In FSE it is constantly ~0.160 ms and in Borderless Fullscreen it is constantly ~0.080 ms. I know this number normally fluctuates somewhat but after a few hours of playing while keeping on eye using both FSE and Borderless Fullscreen it's very obvious to me that FSE close to double vs Borderless Fullscreen which I found fairly easy to notice since in Borderless Fullscreen it very rarely goes over 0.100ms and in FSE it is always well over 0.100ms and constantly hovers around 0.160ms.

joeygreco1985
Posts: 1
Joined: 14 Jan 2022, 01:11

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by joeygreco1985 » 14 Jan 2022, 01:13

Hey has there been any updates on this? I just discovered my PC is doing the same thing and it's driving me insane. Is there a specific driver I can roll back to?

yamaci1775
Posts: 17
Joined: 27 Jan 2019, 02:29

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by yamaci1775 » 22 Jun 2022, 17:30

having the exact same issue as of 516.40 driver... :(

Kuscheln_Hammer92
Posts: 62
Joined: 29 Apr 2023, 08:57

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by Kuscheln_Hammer92 » 01 May 2023, 16:00

I think this is what I'm experiencing on my amd card actually. I'm trying to play halo MCC on fixed refresh rate for my backlight strobing. Enabling RTss causes much worse performance in fps and frametimes than simply using game fps limiter.

User avatar
Chief Blur Buster
Site Admin
Posts: 11653
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by Chief Blur Buster » 11 May 2023, 01:53

Kuscheln_Hammer92 wrote:
01 May 2023, 16:00
I think this is what I'm experiencing on my amd card actually. I'm trying to play halo MCC on fixed refresh rate for my backlight strobing. Enabling RTss causes much worse performance in fps and frametimes than simply using game fps limiter.
Fixed refresh rate does not LFC.

LFC is only used when VRR is turned on. (VRR is not fixed refresh rate).

Can you explain better? Your problem may be completely something else entirely, if you actually meant a non-VRR fixed refresh rate.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

bumbeen
Posts: 86
Joined: 25 Apr 2023, 14:35

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by bumbeen » 11 May 2023, 04:00

have u guys tried using nvidia frameview to record the individual frame present times? I only play valorant so I can only speak to that game, but I get much more consistent frame times on my 270hz with g-sync + vsync with a 251fps cap than a 265 fps cap. The OSD will show refresh rate all over the place, but .01% lows are much better with 251 cap than 265.

Here is a sample with fps cap at 265 with g-sync vsync on, first column is MsBetweenPresents, second column is MsBetweenDisplayChange:

3.741,3.756
3.722,3.757
3.556,3.514
3.771,3.761
3.919,5.709
8.875,7.110
0.846,0.994
2.000,2.042
3.484,3.276
3.522,6.120
4.265,1.658
3.288,3.308
3.944,4.012
3.825,3.792
3.704,3.688
3.710,3.688
3.757,3.753
3.952,5.814
3.672,1.759
3.771,5.673
3.935,2.035
3.577,5.456
3.776,1.903
3.984,5.801
3.666,1.841
3.820,3.835
3.867,3.893
3.775,5.455
3.686,2.033


and with 251 fps cap:

3.128,3.245
4.254,4.156
3.879,4.451
4.190,3.610
3.860,4.087
3.801,3.585
3.924,4.274
4.607,4.377
3.414,3.272
3.876,3.880
4.124,4.271
3.959,3.817
4.027,4.033
3.930,4.050
3.916,3.822
4.313,4.293
3.759,3.717
3.990,4.072
4.142,4.463
3.958,3.631
3.965,4.227
3.978,3.925
3.933,3.701
3.959,3.945
3.922,3.935
4.128,4.098
4.061,4.072
3.808,3.817
3.941,3.936
4.310,4.299
3.933,3.950
4.136,3.955
3.726,4.144
4.012,3.822
4.434,4.341
3.429,3.431
4.265,4.244
3.707,4.147


my theory is frame rate chugs along at 165 ok, then something in game happens, and suddenly frame render takes longer than normal. So now FPS is <165, and engine speeds up to compensate, but can only go 5fps more, and if the drop was more than 5fps, then it still h as to throw more frames to catch back up, so I start getting backpressure on the cpu. With more runway for raising FPS from the 251 cap, the backpressure never happens. But I don't have issue with LFC as the range is 48-270.

Possible there is CRU tweak that can be made to monitor to raise g-sync lower bound or no? Sorry if this isn't helpful, just trying to give some data point :)

Kuscheln_Hammer92
Posts: 62
Joined: 29 Apr 2023, 08:57

Re: Monitor reported refresh rate vs system [Technical LFC Debugging for VRR]

Post by Kuscheln_Hammer92 » 12 May 2023, 15:00

Chief Blur Buster wrote:
11 May 2023, 01:53
Kuscheln_Hammer92 wrote:
01 May 2023, 16:00
I think this is what I'm experiencing on my amd card actually. I'm trying to play halo MCC on fixed refresh rate for my backlight strobing. Enabling RTss causes much worse performance in fps and frametimes than simply using game fps limiter.
Fixed refresh rate does not LFC.

LFC is only used when VRR is turned on. (VRR is not fixed refresh rate).

Can you explain better? Your problem may be completely something else entirely, if you actually meant a non-VRR fixed refresh rate.
We can turn this into PMs if you'd prefer. Whenever I don't have RTSS enabled in any fashion besides monitoring stats, I experience drastic cuts to performance and utilization. Even without passive CPU enabled. Using the rivatuner framerate limiter or using scanline sync it prefers to instead just the framerates to half the refresh rate of the monitor rather than match it. This is when I'm pulling 200+ fps without them on. So I reconciled to using enhanced sync and letting framerates go above the fixed refresh rate but still limiting it with the in game limiter. The in game limiter so that it's not bouncing around a whole bunch and just having a nice level ceiling. Honestly, enhanced sync with the in game limiter was awesome. It was a grunt birthday bananza. lol

Post Reply