PC Microstuttering

4k - reflection quality Ultra+ (no other changes to the previous screenshot)

52fps and a frametime that leaps between 8ms to near 30ms, it feels awful, juttering, jarring, microstutter.

Again not about fps, i've dropped 2fps and it feels completely trash.





1080p - reflection quality Ultra+ (no other changes to the previous screenshot)

63fps and a frametime that leaps between 8ms to near 30ms, it feels awful, juttering, jarring, microstutter.

Again not about fps, my gpu usage has dropped to 40% because it is working at a fraction of the workload and i've gone up 10fps and it feels completely trash.





 
Last edited by a moderator:
Another example but from a custom map so not possible to replicate.

4k and 1080p screenshots, both using Ultra+ reflection quality.

4k = 54fps and feels awful. 1080p = 76fps (49% gpu load) and feels just as bad.









 
Is it micro stutter (several times per second) or just freezes every minute or when moving into a new area (loading/spawning)?Is vsync on or off?

Is the monitor at 60hz or one of the high refresh models?

Are you using nVidia's fast or adaptive vsync or triple buffering or gsync?
Click to expand...
The freeze seems to be always when I'm fighting. Vsync is off, 60hz monitor running at 1600x900 for it helps with fps. I have a 1080p monitor. No on nvidia's vsync / triple buffereing.

 
I have a feeling most people don't know what Microstuttering is. So you might have to clearly define it first, before you make a thread like this.

The general acknowledged definition is a highly variable frametime (not frames per second) causing jankiness/stuttering because one frame is rendered faster than the next. PCGamer had an interesting article about it: https://www.pcgamer.com/what-is-microstutter-and-how-do-i-fix-it/

Edit: +1 to SurvivalUK's posts. What tool are you using there to measure frametime?

 
I've never used gsync, but I would think it requires vsync on, since it is matching the vsync to the game's FPS.
Click to expand...
Nope, if anything, we G-Sync users prefer to disable V-Sync and go the extra mile by simply limiting fps through other means like Precision XOC which allows for maximum fps up to the limit in games instead of rounding down which is what V-Sync does.

But this has been around for a long time, long before I got my G-Sync monitor and was using v-sync (tried both on and off). But I know that's irrelevant info since you guys don't care about previous game versions.

I have a feeling most people don't know what Microstuttering is.
Click to expand...
More ppl know about it than you think.

https://7daystodie.com/forums/showthread.php?25024-Movement-stutter-frame-lag-while-at-60-fps

 
Last edited by a moderator:
Xeon 3.4ghhz gtx 1060 6gb ssd 24gb ram .Very few stutters up to 17.1 now full 1-2 second stutters every 15-20 mins and one full lock up opening a tower door in wooden trader fort.

 
Nope, if anything, we G-Sync users prefer to disable V-Sync and go the extra mile by simply limiting fps through other means like Precision XOC which allows for maximum fps up to the limit in games instead of rounding down which is what V-Sync does.But this has been around for a long time, long before I got my G-Sync monitor and was using v-sync (tried both on and off). But I know that's irrelevant info since you guys don't care about previous game versions.

More ppl know about it than you think.

https://7daystodie.com/forums/showthread.php?25024-Movement-stutter-frame-lag-while-at-60-fps
Click to expand...
Exactly, I always play with vsync off and cap my fps 3-4 below my monitors minimum refresh rate. However, for whatever reason in 7DTD having vsync off was resulting in microstutter regardless.

 
X8 versus X16 will not cause stutter. Look at tests running cards in X8 versus X16. So far no game can saturate X8, so it was likely due more to being loose than not enough bandwidth. In fact one thing you can do when low on PCI-E lanes is to force both video cards into X8 assuming you run SLI and need more lanes for NVME or other PCI-E stuff. No difference in framerates in X16/X16, X16/X8, or X8/X8 during testing done this past year.

I wanted to respond to the garbage collection though. I normally get the regular stuttering after playing a while (1hr or more?) and loading into a new area, so this makes sense to me.

Also, Fox, anybody not using V-Sync has no clue how it interacts with G-Sync. Again, just because the sheople do it religiously, doesn't make it right. Vertical sync prevents tearing due to pumping out incomplete frames or attempting to pump out more frames than the refresh rate of your monitor allows. G-Sync is the opposite of that. It adjusts the refresh rate down to your framerate when your rate drops below the monitor refresh rate.

For example, assume a 60Hz monitor. You're in the desert and you can pump 90fps out. VSync matches full frames to your refresh rate and prevents tearing. As you move into the burned forest you drop to 45fps. G-Sync would then lower your monitor refresh rate to 45Hz to prevent stuttering. Both should be on.

 
Last edited by a moderator:
X8 versus X16 will not cause stutter. Look at tests running cards in X8 versus X16. So far no game can saturate X8, so it was likely due more to being loose than not enough bandwidth. In fact one thing you can do when low on PCI-E lanes is to force both video cards into X8 assuming you run SLI and need more lanes for NVME or other PCI-E stuff. No difference in framerates in X16/X16, X16/X8, or X8/X8 during testing done this past year.
I wanted to respond to the garbage collection though. I normally get the regular stuttering after playing a while (1hr or more?) and loading into a new area, so this makes sense to me.

Also, Fox, anybody not using V-Sync has no clue how it interacts with G-Sync. Again, just because the sheople do it religiously, doesn't make it right. Vertical sync prevents tearing due to pumping out incomplete frames or attempting to pump out more frames than the refresh rate of your monitor allows. G-Sync is the opposite of that. It adjusts the refresh rate down to your framerate when your rate drops below the monitor refresh rate.

For example, assume a 60Hz monitor. You're in the desert and you can pump 90fps out. VSync matches full frames to your refresh rate and prevents tearing. As you move into the burned forest you drop to 45fps. G-Sync would then lower your monitor refresh rate to 45Hz to prevent stuttering. Both should be on.
Click to expand...

Well in my case, it wasn't just limiting it to x8. If it was, I would have probably never known and ended up with something like power issues or something even worse. I actually used a version of kombuster that is built into GPU-Z to test what PCI-E mode it was using. It never went from 1.0 before I reseated the card. After, it went to x16 instantly and I never had the problem again.

Also, I am not suggesting that people are having bandwidth issues and that is what is causing it. Mostly, I think there is a problem with certain configurations of parts and drivers people are using which is exposing edge cases. Turning on V-Sync is a suggestion to see if that would help even out frame times. It seems to work for most people. It is not at all surprising to me that adaptive sync users would also suffer from frame pacing issues.

Another thing to note, enabling V-Sync in D3D games will not reduce your frame rate per a mathematical factor of your refresh rate like it used to in the 90's. D3D employs a dirty flip cue to keep this from happening. While it is not triple buffering, which is the correct way to handle that situation, it works to some extent. The remaining problem being mouse lag in certain situations, since this "dirty flip cue" does not always show you the most recent frame available.

I don't personally have any problems with mouse lag in this game with V-Sync enabled. YMMV.

 
Here is what i found after extensive testing and i will say this. A17.1 performance has improved vs A17 for my issue, it's not fixed or great but better at least.
4k - reflection quality low = fps 50's and good frame time

4k - reflection quality ultra+ = fps 50's / p*ss poor frame time (9ms - 30ms range, fps does not change)
Click to expand...
Looks like it is reflections doing it then. No surprise since they are expensive. They have to do 6 renders to update the cubemap. The higher the setting, the more often it updates and the higher the resolution and the renders use a longer distance. At high or better it is doing a lot of extra rendering, which would give you spikes.

I've tweaked reflection settings several times now and nothing will make them faster other than reducing values, which is what the lower settings do.

 
Xeon 3.4ghhz gtx 1060 6gb ssd 24gb ram .Very few stutters up to 17.1 now full 1-2 second stutters every 15-20 mins and one full lock up opening a tower door in wooden trader fort.
Click to expand...
That is not microstutter. That could be garbage collection, but that normally happens about every minute. Could be windows doing some heavy background work or some game system stalling out for a moment like a big load/save.

 
I have a lot of Microstuttering, at worst the stutters can go up to 300ms

Windows 10 64bit fully up to date.

Ryzen 7 1800x / H80i V2 Liquid Cooled

GTX 1060 6gb Palit

16gb 2600mhz DDR4 Corsair Veangance ram

Running from a SSD: Agate 960gb with 550mb/s read 530mb/s write.

Motherboard: ASUS PRIME B350

Vsync ON

Im fine when traveling fast, it seems to happen a lot more to me when im trying to engage in combat.

Besides this issue im running on balanced graphics, 1080p, have 60fps and temperatures are pretty cool. (god bless the 1060)

I am currently running the experimental beta.

 
Last edited by a moderator:
Thanks fo your advices and thanks Faatal to follow this thread. I tried everything suggested by other players, but it didn't change anything. I'll post a video because as it was said it's possible that we don't all talk about the same thing when saying "microstuttering". I have to say it's absolutely not game breaking, just a little annoying, but I can live with it.

However, I killed today a lot of process in the task manager and it gets better. This f***ing Skype that I've no use for seemed to be one of the possible issues. I'll try tomorrow to eliminate an other bundle of process to see how it works.

(My config - ssd - 16g ram - nvdia gtx1060 6g - Ryzen 5 2600 6c)

 
Last edited by a moderator:
The latest build of experimental seems to have fixed the microstuttering for me. For the most part anyway. Sometimes I feel like I can spot a little bit for a brief moment but unsure(maybe placebo). There is definitely one scenario though where it still happens: standing in the space a door occupies.

I started a new game just to see if this build fixed this dreadful microstutter issue plaguing people and noticed whenever I stand in the block space a door occupies, I have microstutters moving through that block space. It doesn't matter if the door is open or closed. I only tested the one POI but it was happening with every door the building had.

Overall, definitely playable now in regards to the microstutter.

 
microstuttering in A17.1 is worst than A17 for me.

I was outside at the roof of a POI. The game suddenly spawn like 10 zombies from a POI in door. The stuttering begin. I can hear them hiting walls and doors. So I turn around to leave the place. 3 Secs later I turn around again and the zombies are all outside. The stuttering never stop and I die right there. I think the game can glitch zombies though walls like noclip when players are not watching. I have seen dogs do that at the same POI. The place has two dogs and 2 zombies at the first level, and when they awake and start hitting the same block, 1 of the dogs always glitched though the wall.

My PC:

i7 6700K

16GB ram, SSD, GTX970,

 
microstuttering in A17.1 is worst than A17 for me.I was outside at the roof of a POI. The game suddenly spawn like 10 zombies from a POI in door. The stuttering begin. I can hear them hiting walls and doors. So I turn around to leave the place. 3 Secs later I turn around again and the zombies are all outside. The stuttering never stop and I die right there. I think the game can glitch zombies though walls like noclip when players are not watching. I have seen dogs do that at the same POI. The place has two dogs and 2 zombies at the first level, and when they awake and start hitting the same block, 1 of the dogs always glitched though the wall.

My PC:

i7 6700K

16GB ram, SSD, GTX970,
Click to expand...
That doesn't sound like microstutter. It sounds like lag.

 
Except it was a singleplayer game.
Click to expand...
You can lag in single player. Somebody more clever can confirm, but as I understand it you get graphics lag and network lag. Perhaps yours was graphics lag. I've had a few lag spikes myself like you describe, though less severe, and I play on a multiplayer server mostly.

The microstutter is a constant jittering that you can see when you look at an object and strafe around it. Try this with any quality settings but make sure reflections are on, and you will see what I mean.

 
Lag refers to network latency, not framerate. Stuttering, choppiness, and bad video performance is just that. The "graphics lag" you speak of is the video device working at capacity and being unable to keep up, thus causes stuttering, loss of framerate, etc. Bad framerate is NOT lag. It's like calling my pickup a tank. I can do it, but it doesn't mean I drive a tank.

Also, I have everything maxed on my GTX 1070 in 1440p without the stutter. What fixed it on my end was removing a registry key which, when created, caused G-Sync to fail and the stuttering you describe here. There are two fullscreen registry entries. One is fine and the other causes a whole mess of issues. My solution was to go into the game, set everything how I wanted it, exit the game, and then delete this one key. Now everything works and no more stuttering. The key I removed is "Screenmanager Is Fullscreen mode_h3981298716". There's also a normal fullscreen key which handles fullscreen. Not sure what this does, but it killed my framerate, disabled G-Sync, and introduced the micro-stutter.

If you want to test what removing the key would do, just rename it and add ".bak" to the end of the name. This way if it doesn't work for you, you just rename it back and you're golden.

 
You must log in or register to reply here.
Back
Top