Micro-stuttering / hitch & Frametime spikes issue

[jorimt]

None of the threads is maxed out? Could it be bottlenecked even not reaching 85% usage?

I'm going to try downsampling and I'll give you some feedback.

Thread percentage doesn't necessarily rule out CPU bottleneck, so yes, downsample, and see if it helps.

[rafinesse]

I think I fixed the frametime spikes for me personally by turning off v-sync globally after having read that suggestion on reddit. My C9 is Gsync compatible 40-120, so I cap fps at 116 and don't aim to go below 40 anyway.

do you consider the frametime graph I captured here normal? or is there anything unusual? thanks in advance for any input.

phpBB [video]

[jorimt]

do you consider the frametime graph I captured here normal? or is there anything unusual? thanks in advance for any input.

Your frametime performance looks (more than) fine. I don't see an issue there at all.

[rafinesse]

Thanks for the quick reply! I'm afraid my problems persist though. Yesterday I did a clean Windows reinstall with all the necessary mobo drivers and updates as well as updating the bios before. I'm not doing any overclocking, only my RAM is set to XMP1 (tried XMP2 as well).

Here is my rig in userbenchmark. It's an EVGA RTX 3080 Ultra Gaming with the OC switch on: https://www.userbenchmark.com/UserRun/35053865

Ram is Ballistix 2x8GB@3200Mhz.

Here is some evidence. Regarding Control, I was thinking those spikes are due to streaming assets, but I'm not so sure anymore. The spike occurs approx. 10 seconds after every game start and mostly when enemies appear. sometimes without any clear cause though as well. Regarding Fortnite, I can't believe the game runs THIS bad. Something must be wrong.

phpBB [video]

phpBB [video]

I tried almost everything mentioned in this thread, nothing helps. Just ordered a new PSU as I bought this one used and it might be 10 years old. Hope this is the key.

And btw, Forza Horizon 4 is more or less the only game not having any hiccups, if that helps analyzing the problem.

Thanks for any help!

[jorimt]

Regarding Control, I was thinking those spikes are due to streaming assets, but I'm not so sure anymore. The spike occurs approx. 10 seconds after every game start and mostly when enemies appear.

Again, I don't see anything unusual there. It's not uncommon for a game to load more assets at the beginning of each session as it intially caches data.

Regarding Fortnite, I can't believe the game runs THIS bad. Something must be wrong.

Fornite is unfortunately known for sporadic and unreliable performance from session to session, as it is an online game with netcode dependencies. Search it on this forum and you'll find a few threads with the same complaints, and very few solutions.

You could run several sessions in a row in that game, and all would produce wildly differing frametime graphs, some with little to no stutter, and some with tons.

I tried almost everything mentioned in this thread, nothing helps. Just ordered a new PSU as I bought this one used and it might be 10 years old. Hope this is the key.

Not saying this is the case in every game you've played, but sometimes it's just the game. That said, here's hoping your new PSU affects your situation positively.

[rafinesse]

Thanks again! After reading your post I was about to leave it and accept that the stutters are related to game engine/asset streaming.

Then I started Cloudpunk. Annoyed by the occasional stutter while rotating the camera, I found a thread where you guys helped the developers to deliver a smooth experience. The result was a "butter smooth" update released in July.

https://steamcommunity.com/app/746850/d ... 291949710/

Now, are these stutters I recorded still streaming/engine related or is Cloudpunk really supposed to be butter smooth as mentioned in the thread above?

[spoiler]

phpBB [video]

phpBB [video]

[/spoiler]

Second video is capped to 45FPS using RTSS as someone else mentioned if it still stutters capped at 45FPS, the problem is related to Motherboard, CPU or RAM. is this true?

Changing the PSU, unfortunately and probably as expected, didn't help.

Is there anything else I can try or should I buy a new Motherboard, CPU and RAM to test all the components? Or is this still asset streaming/engine related?

Other fixes I tried:
- played with only 8 GB RAM, tried both of them separately
- capped CPU clock to 3600Mhz as I had the feeling stutter was due to CPU boosting
- low latency off/on/ultra
- energy mode AMD Ryzen Balanced/High performance/Windows High performance
- checked nvme ssd performance and temperature

If someone could test Cloudpunk, preferably with an RTX 3080/3090, and check for these stutters, that would be highly appreciated.

Edit: According to this post, it might just be a driver issue: https://www.nvidia.com/en-us/geforce/fo ... with-3080/

[EeK]

do you consider the frametime graph I captured here normal? or is there anything unusual? thanks in advance for any input.

Your frametime performance looks (more than) fine. I don't see an issue there at all.

I'm honestly kind of surprised that @rafinesse is getting such stable frametimes with VSync disabled both in-game and at the driver level (and, apparently, without an FPS cap - but they wouldn't be hitting the C9's max 120Hz in Control at 4K with RT on, even with a 3080 and DLSS rendering at 1440p).

@jorimt, shouldn't they be using G-Sync combined with VSync on?

Also, you mentioned that UE3 is bad with streaming assets in open world titles, guess UE4 is the same, as I've been playing Jedi: Fallen Order, and getting constant hitches/stutters when loading new areas/streaming assets - and that's not even a true open world game! They don't even do a good job at hiding the texture streaming, as you can see loads of pop-in for all sorts of assets in the scenery.

[rafinesse]

Someone told me to try Doom Eternal as it is said to be perfectly optimized - and indeed I had far less frametime spikes, just one in ten minutes of playing (refunded right away, not my kind of game). Got my hopes back that I don't have any faulty parts.

I'm pretty disappointed in PC gaming though as it seems pretty hard to find optimized games these days. In most games I tested, even tweaking the settings didn't help (Cloudpunk i.e. still has spikes @1080p medium - not to mention Watch Dogs Legion). Sure, I was looking for errors, but it's been pretty obvious most of the time.

@EeK I think I was just imagining things - probably no difference between V-sync on and off. As I'm in the C9's Gsync range, V-sync shouldn't make a difference anyway, right?

[jorimt]

Thanks again! After reading your post I was about to leave it and accept that the stutters are related to game engine/asset streaming.

Then I started Cloudpunk. Annoyed by the occasional stutter while rotating the camera, I found a thread where you guys helped the developers to deliver a smooth experience. The result was a "butter smooth" update released in July.

There's sync-induced stutter and there's system-induced stutter. G-SYNC fixes the former.

System-side stutter due to frame pacing methods at the engine-level can be addressed by the devs in certain circumstances to allow for smoother G-SYNC interaction, but any stutter caused by asset loading and/or CPU-related issues will remain.

If you didn't know, a frametime "spike" represents a single frame render that lasts longer than a single refresh cycle. This means it can skip one or more frame delivery cycles as it completes, forcing the GPU to repeat the previous frame one or more times until it is finally ready for display.

G-SYNC itself can only aligns these frames once they're available, and has no control over when they are; that's up to the system.

If someone could test Cloudpunk, preferably with an RTX 3080/3090, and check for these stutters, that would be highly appreciated.

I'd test it, but I have neither.

Edit: According to this post, it might just be a driver issue: https://www.nvidia.com/en-us/geforce/fo ... with-3080/

Possibly.

I'm honestly kind of surprised that @rafinesse is getting such stable frametimes with VSync disabled both in-game and at the driver level

Having V-SYNC off will not affect the severity of frametime spikes. Those are independent of syncing methods. It will tear and stutter instead of just stutter in those instances though.

@jorimt, shouldn't they be using G-Sync combined with VSync on?

For tear-free G-SYNC yes, but it's up to the user to decide.

Also, you mentioned that UE3 is bad with streaming assets in open world titles, guess UE4 is the same, as I've been playing Jedi: Fallen Order, and getting constant hitches/stutters when loading new areas/streaming assets - and that's not even a true open world game! They don't even do a good job at hiding the texture streaming, as you can see loads of pop-in for all sorts of assets in the scenery.

I own and played through it at launch. It's certainly a hard drive thrasher. This also affected consoles. It's the game; some UE4 games handle this better than others, but, no, UE4 doesn't seem to have a great streaming system either.

[jorimt]

As I'm in the C9's Gsync range, V-sync shouldn't make a difference anyway, right?

As per my G-SYNC 101 Closing FAQ:
https://blurbusters.com/gsync/gsync101- ... ttings/15/

Wait, why should I enable V-SYNC with G-SYNC again? And why am I still seeing tearing with G-SYNC enabled and V-SYNC disabled? Isn’t G-SYNC suppose to fix that?

The answer is frametime variances.

“Frametime” denotes how long a single frame takes to render. “Framerate” is the totaled average of each frame’s render time within a one second period.

At 144Hz, a single frame takes 6.9ms to display (the number of which depends on the max refresh rate of the display, see here), so if the framerate is 144 per second, then the average frametime of 144 FPS is 6.9ms per frame.

In reality, however, frametime from frame to frame varies, so just because an average framerate of 144 per second has an average frametime of 6.9ms per frame, doesn’t mean all 144 of those frames in each second amount to an exact 6.9ms per; one frame could render in 10ms, the next could render in 6ms, but at the end of each second, enough will hit the 6.9ms render target to average 144 FPS per.

So what happens when just one of those 144 frames renders in, say, 6.8ms (146 FPS average) instead of 6.9ms (144 FPS average) at 144Hz? The affected frame becomes ready too early, and begins to scan itself into the current “scanout” cycle (the process that physically draws each frame, pixel by pixel, left to right, top to bottom on-screen) before the previous frame has a chance to fully display (a.k.a. tearing).

G-SYNC + V-SYNC “Off” allows these instances to occur, even within the G-SYNC range, whereas G-SYNC + V-SYNC “On” (what I call “frametime compensation” in this article) allows the module (with average framerates within the G-SYNC range) to time delivery of the affected frames to the start of the next scanout cycle, which lets the previous frame finish in the existing cycle, and thus prevents tearing in all instances.

And since G-SYNC + V-SYNC “On” only holds onto the affected frames for whatever time it takes the previous frame to complete its display, virtually no input lag is added; the only input lag advantage G-SYNC + V-SYNC “Off” has over G-SYNC + V-SYNC “On” is literally the tearing seen, nothing more.