FPS issues with VR (Oculus Rift)

Hello!

I've seen people saying (braging?) they have maxed out settings in DCS with their 1070 GPU while using Oculus Rift, and I can't even come close with my 2070 GPU.

What is wrong with my system?

Specs:

CPU : Intel Core i7-5930K 4,5 GHz 6 cores

RAM : 16GB DDR4 2666MHz

GPU : MSI RTX 2070 Gaming Z 8 GB

OS : Windows 10 Professional

DCS version : 2.5.6 Open Beta

Here are a few screenshots of the performance monitor while playing a single player mission (instant action - free flight - Huey - Caucasus)

CPU:

RAM:

GPU:

And my DCS settings:

Oculus Tray Tool settings:

And my frame rate in-game:

I don't need fancy graphics, but what I want is solid FPS at 45 (max FPS for Oculus Rift), and SHARP images

Any advice greatly appreciated.

Hello!

 

I've seen people saying (braging?) they have maxed out settings in DCS with their 1070 GPU while using Oculus Rift, and I can't even come close with my 2070 GPU.

 

What is wrong with my system?

 

Specs:

CPU : Intel Core i7-5930K 4,5 GHz 6 cores

RAM : 16GB DDR4 2666MHz

GPU : MSI RTX 2070 Gaming Z 8 GB

OS : Windows 10 Professional

DCS version : 2.5.6 Open Beta

 

Here are a few screenshots of the performance monitor while playing a single player mission (instant action - free flight - Huey - Caucasus)

 

CPU:

 

RAM:

 

GPU:

 

 

And my DCS settings:

 

Oculus Tray Tool settings:

 

And my frame rate in-game:

 

I don't need fancy graphics, but what I want is solid FPS at 45 (max FPS for Oculus Rift), and SHARP images

 

Any advice greatly appreciated.

First off...

1. Do you use Nvidia control panel?

2. What is your DCS VR PD set to? If you’re using OTT, make sure the DCS PD is set to 1.0.

Other than those two questions, the only thing that I see for FPS vs. sharpness is AF. AF doesn’t hit FPS too bad but PD does. AF helps a lot with jiggly lines. I would try 16x AF, shadows to flat, and set your OTT PD just below the stutter point. Mine is at 1.2. If you have the Rift S you will only get 40 FPS due to 80hz. 45 FPS is for CV1 @ 90hz. Using ASW at this point is your best option or FPS will bounce around a lot, sometimes making it tough to look at.

Turn off SSAA . It does not work for VR . Also , i see you are running pd 1.6 . That is way to high for a 1070 .

Great advice already given. I'll just add:

Set textures to High. Medium make the cockpit look terrible.

Shadows are expensive. Try turning them off and see if you can live without them.

When benchmarking CPU always look at the individual cores. There should be one that does most of the work.

Maybe take a look at my post here (and the link within) and try to replicate what I do and see if you get similar results. I don't know how close a 2070 is to a 1080 ti, but I'm guessing pretty close. If the results are very different, then there might be an issue.

Note what I'm doing with the nvidia settings - someone recommended this to me on here - it improves AA at less cost that going 4x in DCS.

I use a PD of 1.2. Set it in DCS, and (at least for testing) don't use OTT at all.

Unfortunately, getting decent frame rates in VR is getting harder and harder with each DCS main engine update. You should be able to get 40 fps easily, but settings have to be set fairly low.

Turn off SSAA . It does not work for VR . Also , i see you are running pd 1.6 . That is way to high for a 1070 .

I have a 2070 card, but was using people with 1070 as an example.

I'll turn off SSAA to if that helps.

Great advice already given. I'll just add:

 

Set textures to High. Medium make the cockpit look terrible.

 

Shadows are expensive. Try turning them off and see if you can live without them.

 

When benchmarking CPU always look at the individual cores. There should be one that does most of the work.

 

Maybe take a look at my post here (and the link within) and try to replicate what I do and see if you get similar results. I don't know how close a 2070 is to a 1080 ti, but I'm guessing pretty close. If the results are very different, then there might be an issue.

 

Note what I'm doing with the nvidia settings - someone recommended this to me on here - it improves AA at less cost that going 4x in DCS.

 

I use a PD of 1.2. Set it in DCS, and (at least for testing) don't use OTT at all.

 

Unfortunately, getting decent frame rates in VR is getting harder and harder with each DCS main engine update. You should be able to get 40 fps easily, but settings have to be set fairly low.

Thanks! I'll have a look! :)

Another suggestion to get some more fps is Kegetys's mod:

https://forums.eagle.ru/showthread.php?t=215373

It does introduce some glitches, but can make a considerable difference.

I had always used it since he released it, but hadn't been using it for the past few months because I've mostly been flying very simple missions, and was scraping by. It also involves a minor bit of installation effort, and it has to be reapplied for every update.

Anyway, just applied it again, and tested. For me it means that for the test I described in my previous post, I can stay at 80 fps for the entire flight. Notice how GPU % always stays below (but close to) 100%, so ASW never needs to kick in.

Thanks! That looks promising, but I hate custom mods, and all those multiplayer integrity problems which follows. I can't believe why ED hasn't implemented this as standard, because it obviously will help a lot on the performance, as I can read from other users in the thread.

I will keep it in mind, though!

Thanks! That looks promising, but I hate custom mods, and all those multiplayer integrity problems which follows. I can't believe why ED hasn't implemented this as standard, because it obviously will help a lot on the performance, as I can read from other users in the thread.

 

I will keep it in mind, though!

ED hasn’t implemented it because it destroys some of the nice graphic effects that they worked hard to produce, and secondly, not everyone sees a benefit with this mod. Mostly lower end systems show some benefit.

Speaking purely personally , i did not like it at all , even on my lower-end system .

ED hasn’t implemented it because it destroys some of the nice graphic effects that they worked hard to produce, and secondly, not everyone sees a benefit with this mod. Mostly lower end systems show some benefit.

The issue is that the nice graphics ED worked hard to produced have now destroyed VR gaming ... And if you turn things down to have it running it is uglier than ever before...

The issue is that the nice graphics ED worked hard to produced have now destroyed VR gaming ... And if you turn things down to have it running it is uglier than ever before...

Simply untrue.

Thanks! That looks promising, but I hate custom mods, and all those multiplayer integrity problems which follows. I can't believe why ED hasn't implemented this as standard, because it obviously will help a lot on the performance, as I can read from other users in the thread.

 

I will keep it in mind, though!

It's not a magic bullet, and it does introduce a number of glitches, certain effects don't look as good. It depends on what's most important to you. If it's high frame rates with AA, then it is very useful. I find I can get a 25% fps improvement, and it's not because I have a "low end system".

As per Oculus's recommendations, games should be running (Rift S) at 80 fps for the optimal VR experience. ASW and 40 fps is a workaround and compromise. Yet I see many on here stating that they're very happy with the beautiful high graphics settings and getting 20 or 30 fps. As I say, it depends on personal taste (or more cynically, an ability for self-delusion).

At the very least, I always try to achieve 80 fps when it's just my aircraft and nothing else, and accept that ASW will kick in when things get busier. Frame rates below 40 are unacceptable. Although ASW does a remarkable job, the artefacts introduced by it can at times be quite distracting.

As to AA. I'm quite impressed that with it turned off things don't look as bad as they used to. Unfortunately, the jaggies from scenery objects like runways or buildings look dreadful. The cockpit surprisingly almost always looks acceptably good without AA - until you get shiny edges, when the shimmering jaggies look atrocious - this I just can't accept. Is there no way to (optionally) dial these effects down for the benefit of VR users? I suspect that it's the feature that applies AA to only a central portion of the screen that is responsible for most of the improvement from this mod (perhaps that alone could be implemented as an option by ED?)

Before ED made the change to deferred rendering in the engine I could easily achieve 120 fps (equivalent) with similar settings on a 1070 and CV1. That change cut my frame rates in half. Decades (yes really) of my being an enthusiastic supporter of ED were flushed down the toilet when they took that decision. Since then, every engine improvement takes its toll. 2.5.5 - 2.5.6 has seen a decrease of 15 - 25 % (for me personally).

ED have lost $100s of my custom since the DR change. I had in the past purchased aircraft that I have hardly flown, just to support them. If they want another penny from me, they are going to have improve the VR (w/ AA) fps considerably.

P.S. Seriously, forget about a PD of 1.6. Try 1.2.

P.P.S. Night is noticeably improved in 2.5.6. Night before the deferred rendering change was terrible (unusable). That's to be welcomed. All imho, of course.

I find I can get a 25% fps improvement, and it's not because I have a "low end system".

Out of interest, how do you measure that? I've never really understood it when people say things like that. I mean, surely with ASW (MP on WMR) you only get 45 or 90 fps in my case, 40 and 80 in yours, I think?

Reduction in frame times, yes, but increase in fps? Just curious.

Out of interest, how do you measure that? I've never really understood it when people say things like that. I mean, surely with ASW (MP on WMR) you only get 45 or 90 fps in my case, 40 and 80 in yours, I think?

 

Reduction in frame times, yes, but increase in fps? Just curious.

It used to be easier

(hi again! :) )

I have to estimate it now (and as per your post - it is in practice indicative) using a mixture of: past experience, the frame rate count at the main menu before putting on the HMD, the measurable increase in GPU% util, turning off ASW w/ OTT.

EDIT: I suppose I could say I get 40 fps in 2.5.6, and 80 fps in 2.5.6 w/ the mod, so 100% fps increase, but that would be naughty.

It used to be easier

 

(hi again! :) )

 

I have to estimate it now (and as per your post - it is in practice indicative) using a mixture of: past experience, the frame rate count at the main menu before putting on the HMD, the measurable increase in GPU% util, turning off ASW w/ OTT.

 

 

EDIT: I suppose I could say I get 40 fps in 2.5.6, and 80 fps in 2.5.6 w/ the mod, so 100% fps increase, but that would be naughty.

The only reliable way of monitoring improvements/degradations in performance is to look at frame times, which is most easily done with fpsVR.

Estimates and past experience can be subject to placebo etc.!

The only reliable way of monitoring improvements/degradations in performance is to look at frame times, which is most easily done with fpsVR.

 

Estimates and past experience can be subject to placebo etc.!

Fine. There is only so much effort each of us is willing to put in as unpaid beta testers for ED. You can see from my screenshot that the GPU util is indisputably, measurably lower with the mod (around 20% -i.e. from around 90% to 72% at the start of the mission).

Not perfect, but surely of more use than all the "it looks smoother to me" posts.

P.S. If when I was able get an fps difference measurement in the recent past it equated to a certain GPU util difference, it is not unreasonable to assume a similar relationship today.

Fine. There is only so much effort each of us is willing to put in as unpaid beta testers for ED. You can see from my screenshot that the GPU util is indisputably, measurably lower with the mod (around 20% -i.e. from around 90% to 72% at the start of the mission).

 

Not perfect, but surely of more use than the all of "it looks smoother to me" posts.

 

P.S. If when I was able get an fps difference measurement in the past it equated to a certain GPU util difference, it is not unreasonable to assume a similar relationship today.

Using fpsVR is no effort at all though. Set it to start with SteamVR and it is always there, and you can toggle it off and on with a key press.

Also, not sure why you are using Afterburner to monitor VR performance. fpsVR is king!

Using fpsVR is no effort at all though. Set it to start with SteamVR and it is always there, and you can toggle it off and on with a key press.

 

Also, not sure why you are using Afterburner to monitor VR performance. fpsVR is king!

I'll look into it next time. I'm not using DCS under Steam.

EDIT - it seems to only work with SteamVR, is that right?

You are taking this discussion way off the point, btw. Are you saying that you see no difference when you use the mod? Perhaps you could point me to the results you have posted which demonstrate this?