Reverb high CPU and GPU frame time - UNPLAYABLE

The OP has got other issues going on that no one can figure out. A 1.2 PD equivalent setting for SteamVR SS, i.e. 150%, is not a high setting by any stretch.

Im interrested in your complete settings. Nvidia CP, steamVR and DCS. Preferably print screen pics for easy copying to my settings.

Confirm that your SteamVR reports 100% = 2204 x 2160 pixels, and 150% = 2700 x 2644 pixels?

There have been different numbers for 150%, because of this, the question.

Folloving the Thuds guide at VR4dcs.com guide and steamVR in Beta should give these numbers. If using the other/older version 188% = native and then 150% is quite low.

2700*2644 times two screens isnt that low. If I make the calculation correct it sums up to 5400*2644.

I will help OP online to try to fix the problem.

If OP have a problem and I try to tell him that he should lower the SS due to bad fps, we shouldnt tell him that SS150% is not a problem. Because it could be part of the problem, we do not know that yet. So, generally it should be a bad idea to tell OP that 150% is no problem at this time.

Im interrested in your complete settings. Nvidia CP, steamVR and DCS. Preferably print screen pics for easy copying to my settings.

 

Confirm that your SteamVR reports 100% = 2204 x 2160 pixels, and 150% = 2700 x 2644 pixels?

There have been different numbers for 150%, because of this, the question.

Folloving the Thuds guide at VR4dcs.com guide and steamVR in Beta should give these numbers. If using the other/older version 188% = native and then 150% is quite low.

 

2700*2644 times two screens isnt that low. If I make the calculation correct it sums up to 5400*2644.

 

 

 

I will help OP online to try to fix the problem.

If OP have a problem and I try to tell him that he should lower the SS due to bad fps, we shouldnt tell him that SS150% is not a problem. Because it could be part of the problem, we do not know that yet. So, generally it should be a bad idea to tell OP that 150% is no problem at this time.

Look, this is all a matter of opinion. I would say that an equivalent SS (150%) of 1.2 PD is pretty low. You may disagree, and that is your right.

Yes, your numbers are correct. The 188% thing went out months ago.

Again, in my opinion, one can tinker with DCS/Steam settings for ever, but the only things that make a significant difference to performance are MSAA and SS/PD. Shadows also make a difference, but all the other stuff is largely irrelevant.

In my experience also, things like Project Lasso and disabling hyper-threading etc. are also a waste of time. As is the infamous Shaders Mod - a lot of visual degradation for no real benefit.

All just an opinion based on a long time testing over recent years, and if others feel differently, than that's great!

In my experience also, things like Project Lasso and disabling hyper-threading etc. are also a waste of time. As is the infamous Shaders Mod - a lot of visual degradation for no real benefit.

I havent done any mods at all, no shader mod, no Lasso etc and I rund the 9900KS at 5.2Gghz with HT on(as I use the computer to a lot other software that actually benifit from HT).

Still interrested in your settings, I might learn from them.

I did shoot a short Movie from F18->free flight->Kaukasus: turned around, descened and passed the North bridge end South Batumi on heading 030, altitude 250feet. Enganged ATT-mode on AP and looked straight ahead and to the sides. This will show my fps, GPU and CPU timings. "My DCS-settings" with SS120%. It would be very nice to see the same from you.

Of course, as 60fps youtubevideo can not reproject 45 or 90fps correct, it wont be a smooth video.

The stutter seen can be interpretated as the same all the time, due to 45/90fps on 60fps Movie.

fpsVR is in view also. If I do this run in SS150, the gaming isnt smooth anymore. Even if fpsVR shows loads below 22.2ms

I havent done any mods at all, no shader mod, no Lasso etc and I rund the 9900KS at 5.2Gghz with HT on(as I use the computer to a lot other software that actually benifit from HT).

 

Still interrested in your settings, I might learn from them.

 

I did shoot a short Movie from F18->free flight->Kaukasus: turned around, descened and passed the North bridge end South Batumi on heading 030, altitude 250feet. Enganged ATT-mode on AP and looked straight ahead and to the sides. This will show my fps, GPU and CPU timings. "My DCS-settings" with SS120%. It would be very nice to see the same from you.

 

 

Of course, as 60fps youtubevideo can not reproject 45 or 90fps correct, it wont be a smooth video.

The stutter seen can be interpretated as the same all the time, due to 45/90fps on 60fps Movie.

fpsVR is in view also. If I do this run in SS150, the gaming isnt smooth anymore. Even if fpsVR shows loads below 22.2ms

I'm not sure what you are trying to achieve here comparing your DCS experience and mine, but for the record I have attached my settings currently in use.

Without going through the hassle of uploading videos, I can tell you that doing the same scenario, I get GPU frame times of 13.5-16.5 and CPU of 11-14. This compares to your 13-15 and 9-11. In both cases we get solid 45fps. This is what I would expect going from 120% SS to 150%.

Both are perfectly acceptable figures. So, what does that tell us?

I'm not sure what you are trying to achieve here comparing your DCS experience and mine, but for the record I have attached my settings currently in use.

Thanks.

Off course, I willing to learn from others and their settings to make my gaming experience better if possible.

Wow, I am loving this conversation, thanks guys, there's a lot of valuable information there!

I think that as advised, it's best that I do start at the lowest settings with a closely comparable hardware setup to mine (which there seem to be a few of in this thread) to get a good baseline... Same settings, same map/mission and compare our results before making a determination on the best way forward.

Thanks again to all of you for engaging in this discussion!

"Again, in my opinion, one can tinker with DCS/Steam settings for ever, but the only things that make a significant difference to performance are MSAA and SS/PD. Shadows also make a difference, but all the other stuff is largely irrelevant."

I have to disagree here. From what I've read here in the forums and what I experienced for some years now is that some settings next to MSAA and SS/PD do make a huge difference. Default terrain shadows, visible range ie. Civil traffic is just load on the CPU, and causes stutter on my side. Chimney smoke just tanks GPU, because smoke in general does that alot.

Don't get me wrong here... this community is gold, and I found so much useful knowledge here to get to my sweet spot for good visuals and smooth performance (+45 fps) in years.

I'm on SS 160 % with my reverb atm (but I'm using the shader mod), no MSAA, terrain shadows flat!, cockpit shadows low, civil traffic off, chimney smoke 0, since 2.5.6 visib range medium (had that on high). Beware of trees vis, when you fly low. Turn it down till you have no stutter when looking 90 degrees out of the cockpit.

Without the shader mod I'd lower my SS to 120%, and that's all to get the same performance, for it mainly relieves the GPU.

Just my 2 cents ;)

I see many having some settings too high and fiddling with SS without seeing real benefit.

"Again, in my opinion, one can tinker with DCS/Steam settings for ever, but the only things that make a significant difference to performance are MSAA and SS/PD. Shadows also make a difference, but all the other stuff is largely irrelevant."

I have to disagree here. From what I've read here in the forums and what I experienced for some years now is that some settings next to MSAA and SS/PD do make a huge difference. Default terrain shadows, visible range ie. Civil traffic is just load on the CPU, and causes stutter on my side. Chimney smoke just tanks GPU, because smoke in general does that alot.

Don't get me wrong here... this community is gold, and I found so much useful knowledge here to get to my sweet spot for good visuals and smooth performance (+45 fps) in years.

I'm on SS 160 % with my reverb atm (but I'm using the shader mod), no MSAA, terrain shadows flat!, cockpit shadows low, civil traffic off, chimney smoke 0, since 2.5.6 visib range medium (had that on high). Beware of trees vis, when you fly low. Turn it down till you have no stutter when looking 90 degrees out of the cockpit.

Without the shader mod I'd lower my SS to 120%, and that's all to get the same performance, for it mainly relieves the GPU.

Just my 2 cents ;)

I see many having some settings too high and fiddling with SS without seeing real benefit.

Yep, it's all up to the individual, their hardware and their preferences.

I wasn't talking about civ traffic, smoke etc. These are a given to keep off or ultra low. i was meaning things like visibility, water, clutter, terrain shadows etc. None of these make much difference imo. We could all tinker forever, and frankly, I think a lot of people prefer that to actually flying in DCS!

I must take you up on the Shaders Mod thing. With your hardware, I am amazed that you use it. Occasionally, I reinstall it just to remind myself of why I don't use it. The visual compromises to achieve almost no performance benefit are huge. It is a real shame that ED put so much effort into things like water just to have them stripped out by that mod.

A perfect example is in my post here https://forums.eagle.ru/showpost.php?p=4215542&postcount=44

The shader mod murders behind the scenes a lot of postprocessing stuff. Water, cockpit lighting and reflections, all lose a huge amount of fidelity, so it's a bit hard to know for sure what works and what doesn't.

So, dukman83 and I did some tests last night.

Settings tuned down, DCS run smooth.

Tuned up to my settings but SS still at 100, also smooth gaming in F18 quick mission.

GPU frame times similar to mine. CPU frame times, a few ms higher for dukman( like 9-10 for me and 11-13 for dukman).

Afterwards dukman played multiplayer and after a while the CPU frametime did rocket up to ~40ms. Despite this, fpsVR seems to have reported CPU Maxthread 40%. FpsVR only counts the processorload / processortiming connected to DCS, so there might be some other processes that loads the core that DCS use.

Still left to check the bios settings, that the 9900KS is doing 5Ghz on all cores all the time.

Also to check that no other software loads the CPU cores that DCS uses. In this case, reenabling HT might be good to get more cores to spread other processes on.

Dukman has the Steam version of DCS and it seems like the changes in the default.settings file actually not do affect reprojection. Anyone know if steam have any override or own setting for this?

Dukman has the Steam version of DCS and it seems like the changes in the default.settings file actually not do affect reprojection. Anyone know if steam have any override or own setting for this?

That is a Steam setting! Whatever is set in default.settings is what is applied to all games.

That is a Steam setting! Whatever is set in default.settings is what is applied to all games.

I thought so also, but it didnt seem to make any difference. One possibility is that there is more than one default.settings, and we did change the wrong one. The FPS was fixed to 45fps even when having both GPU and CPU fram time well below the 11.1ms limit, Flying around and looking straight up still hit 45fps. The setting was “motionvector“ and after changing to “auto” still only 45fps.

I also saw posts earlier in this forum where they say that changing settings had no effect.

Do you know what happens if you dont have the “Beta” activated? At vr4dcs it say “you must have the Beta installed before proceeding”

Looks like it says that without the Beta selected/installed you couldnt change these settings ?

I thought so also, but it didnt seem to make any difference. One possibility is that there is more than one default.settings, and we did change the wrong one. The FPS was fixed to 45fps even when having both GPU and CPU fram time well below the 11.1ms limit, Flying around and looking straight up still hit 45fps. The setting was “motionvector“ and after changing to “auto” still only 45fps.

 

I also saw posts earlier in this forum where they say that changing settings had no effect.

Do you know what happens if you dont have the “Beta” activated? At vr4dcs it say “you must have the Beta installed before proceeding”

Looks like it says that without the Beta selected/installed you couldnt change these settings ?

If the setting wasn’t taking effect, you would see fps varying all over between 45 and 90. You will always see 45 with ‘motionvector’, and almost always with ‘auto’ as well. In my experience you have to be below 10 to get it show 90.

I keep mine at ‘motionvector’.

Don’t regard vr4dcs as some kind of bible. Good advice on the whole though.

If the setting wasn’t taking effect, you would see fps varying all over between 45 and 90. You will always see 45 with ‘motionvector’, and almost always with ‘auto’ as well. In my experience you have to be below 10 to get it show 90.

 

I keep mine at ‘motionvector’.

 

Don’t regard vr4dcs as some kind of bible. Good advice on the whole though.

The first setting off course had made effect earlier but the change to auto did not seem to have effect. I Think we where well below 11.1ms ( dont remember the exakt numbers) with the nose pointing up and seeing only blue sky.

The first setting off course had made effect earlier but the change to auto did not seem to have effect. I Think we where well below 11.1ms ( dont remember the exakt numbers) with the nose pointing up and seeing only blue sky.

My apologies, I thought you were suggesting that the MR settings weren't working. Personally, I can't see the point in using Auto. Why do you want the fps to jump to 90 on very rare occasions when looking at sky?

My apologies, I thought you were suggesting that the MR settings weren't working. Personally, I can't see the point in using Auto. Why do you want the fps to jump to 90 on very rare occasions when looking at sky?

I have 90 forv lot more time when looking up in the sky. This was a test to see if OP system was able to produce 90fps, when GPU and CPU framtime is low enough. As fpsVR not do count other software that may consume time thus the total fram time could be more than shown in fpsVR, it should be a good indicator i something else is disturbing.

I just flew a own multiplayer mission as the host and fot the over sea part and all high altitude parts, I had 90fps.

I enjoy 90fps as more smooth than reprojection at 45. So for me auto is the natural option.

BTW, In my case I get 90fps as soon as GPU/CPU comes below 11.1 -just tested it, to be sure.

I do only need to get down to 11.1, not lower. And it happens at the same time, it do not take time to change either.

OK, you're getting down to personal preference and compromises now to reach 90fps.

I think you've got it under control now, so I'll leave you to it!

OK, you're getting down to personal preference and compromises now to reach 90fps.

 

I think you've got it under control now, so I'll leave you to it!

In this case its more about OP’s issues.

So setting ”auto” would be preferred to see that it can switch to 90fps when GPU/CPU frametime is low enough.

I think next step is to rem out the reprojection to see if there is response to the setting. If it is, it might be a question about the core DCS uses for its main thread is loaded (down) with other processes. It might be a rest from fiddeling with project lasso etc.

Also, OP should make sure the BIOS is set so the CPU does 5ghz on all cores all the time.