[REPORTED]VR Performance OB 2.5.6

[imacken]

it is? I didnt see any improvement?

 

Same here, no improvement at all. In fact, I believe it is significantly worse. I am becoming more and more worried about VR and DCS.

[TOViper]

I would love to know your settings!

 

https://forums.eagle.ru/showpost.php?p=4379868&postcount=4

[Likkah]

i spent a lot of money for that game.... but now i cant play :( looks shit with much lower settings as before...

[Gryzor]

It looks like to me that nobody in ED has a HMD or minimum interest in VR, it seem even they don´t test anything. Looking for other issues like many objects not rendered in one eye. I´m very disappointed, they ruined VR experience in DCS the last six months (since 2.5.6).

 

At least I had the other WWII sim wich I can reach 90 fps easily with a superb level of detail, deferred shading, glass dynamic reflections, etc...

[fab.13]

If anyone from ED is reading this: whatever you guys are doing to improve VR performance, is working!

Based on what I've seen so far in 2.5.6.50726, my framerate is quite a bit smoother than in the previous builds (and they were in turn better than the ones before those).

 

Keep it up! :thumbup:

 

+1

 

my free flight test mission is smoother too,

I reduced gamma setting to reduce 'milk atmosphere'

[vracan]

so weird to some people getting much improved VR performance while others are reporting big loss ?!

this will only make ed's job harder for knowing what's working and whats not

[Hippo]

so weird to some people getting much improved VR performance while others are reporting big loss ?!

this will only make ed's job harder for knowing what's working and whats not

 

Without wishing to appear rude I would be sceptical. Take a comment such as

 

Based on what I've seen so far in 2.5.6.50726, my framerate is quite a bit smoother than in the previous builds (and they were in turn better than the ones before those).

 

What does this even mean? Frame rate can be higher or lower, I'm not sure what is meant by "smoother". How much are "quite a bit" and "better"?

 

When I see a post such as the above, without a description of how the testing was carried out, the DCS settings, and quantifiable results for the versions being compared, I take it with a huge pinch of

salt. In fact, I would go as far as to say that such posts are worthless.

 

Specifically, with 2.5.6 there have been (suggested) pathfinding issues, where fps can vary hugely depending on the mission construction.

 

With my own, basic and somewhat flawed testing, all I can say is that frame rates took around a 10% hit from 2.5.5 - > 2.5.6, and have slowly been recovering ever since. This is for a simple F-18 mission with no other objects in it. I have, however, seen missions with an 80% hit in 2.5.6. If you were to read my previous posts on this thread you would see I have at least made a basic attempt to back up my assertions.

 

I suspect that, even though 2.5.6 has gone to stable, many issues remain. Many issues that have been reported on this forum have been allowed to go on to stable without being fixed.

 

Openbeta is now the wild west, with stable being nothing but a figleaf.

 

I have so far not seen a single post where someone has stated that frame rate is greatly improved in 2.5.6, which contains any convincing evidence.

Edited June 11, 2020 by Hippo

[speed-of-heat]

Without wishing to appear rude I would be sceptical. Take a comment such as

 

 

 

What does this even mean? Frame rate can be higher or lower, I'm not sure what is meant by "smoother".

 

When I see a post such as the above, without a description of how the testing was carried out, the DCS settings, and results for the versions being compared, I take it with a huge pinch of

salt. In fact, I would go as far as to say that such posts are worthless.

 

Specifically, with 2.5.6 there have been (suggested) pathfinding issues, where fps can vary hugely depending on the mission construction.

 

With my own, basic and somewhat flawed testing, all I can say is that frame rates took around a 10% hit from 2.5.5 - > 2.5.6, and have slowly been recovering ever since. This is for a simple F-18 mission with no other objects in it. I have, however, seen missions with an 80% hit in 2.5.6. If you were to read my previous posts on this thread you would see I have at least made a basic attempt to back up my assertions.

 

I suspect that, even though 2.5.6 has gone to stable, many issues remain. Many issues that have been reported on this forum have been allowed to go on to stable without being fixed.

 

Openbeta is now the wild west, with stable being nothing but a figleaf.

 

Agreed smoother is subjective, given many people in vr use asw/motion smoothing I’m not sure FPS, is useful. Talking about average frame times for gpu and cpus is better and more precise but even that fails to capture this... for perf we really need a common benchmark we can all use...

 

My own experience for the last few patches, is that my frame times for cpu and gpu are largely unchanged... and are still below the perf I had in 2.5.5, 10% is the right order of magnitude for gpu, but I still see some spikes in cpu times, even though they are mostly good. I also see gpu frame spikes in odd scenarios, like when a jet spools up to launch on the carrier and I am near it, even if I am not looking at it...

 

It’s pretty complex...to be precise when there are , so many variations of perf issue.

Edited June 11, 2020 by speed-of-heat

[Charly314159]

I suspect that, even though 2.5.6 has gone to stable, many issues remain. Many issues that have been reported on this forum have been allowed to go on to stable without being fixed.

 

As somebody said on an other topic we should now call it "Stable Open Beta" because it is at least as borked as the real Beta but with stable bugs ;)

[Hippo]

Agreed smoother is subjective, given many people in vr use asw/motion smoothing I’m not sure FPS, is useful. Talking about average frame times for gpu and cpus is better and more precise but even that fails to capture this... for perf we really need a common benchmark we can all use...

 

My own experience for the last few patches, is that my frame times for cpu and gpu are largely unchanged... and are still below the perf I had in 2.5.5, 10% is the right order of magnitude for gpu, but I still see some spikes in cpu times, even though they are mostly good. I also see gpu frame spikes in odd scenarios, like when a jet spools up to launch on the carrier and I am near it, even if I am not looking at it...

 

It’s pretty complex...to be precise when there are , so many variations of perf issue.

 

Admittedly, I used fps rather lazily, simply because it's a measure everyone instinctively understands. In my case I am using ASW, and my frame rates are mostly fixed at 80 or 40 fps. What does vary is GPU utilisation, and it is that that I had been measuring.

 

My point, though, is that the poster (and many others who post similarly) would be substantially more credible and informative if he had also posted:

 

 

• which mission he'd used to test
  
• how he tested
  
• graphs of util / fps / frame time / any objective measure you like
  
• his DCS settings
  
• his hardware spec (in this case he does, but many don't)
  

 

Surely this is preferable to:

 

Wow, this new version is so much faster.

Edited June 11, 2020 by Hippo

[speed-of-heat]

agreed

[Eldur]

DCS never ran as bad as it does today since 1.2.2 - 1.5.8... complete disaster IMHO.

Edited June 11, 2020 by Eldur

[robgraham]

Frames this latest OB are.. ugly, and not just in the rates, the frame times are a MESS rather then nice smooth frames you can actually watch visibly the frame time spikes happen and seemingly for no reason at all.

 

The Perf is the worst I've seen it in a very long time.

[Dunx]

Sadly, performance has dropped for me too, a lot.

 

I've almost run out of options to disable or degrade, I really don't want to fly in minecraft.

[Dunx]

Oh, did Exports take a hit?

 

I've just clawed back some a load of CPU by disabling Witchcraft.

 

I also run Tacview and VAICOM Pro but they were only about 4% CPU combined, and I can't do without VCP.

[Motomouse]

My Reverb is broken and I wait for the replacement. I use my rift right now while waiting for the Reverb replacement.

 

I am talking about both HMDs:

 

In my opinion the best option to use VR today with currently available hardware:

Get rid of AA, supersampling or pixel density adjustments to improve the image quality.

There are only so much physical pixels on your display and while the rendering at 100% will surly look awful to you at first after a short adaption phase you will stop noticing the shimmering and the jaggies. What do you get in return? Smooth framerates, 60 FPS on the Reverb, 45/90 FPS (motion smoothing) on the Rift, at all times in single player, multiplayer and even on the channel map and on the deck of the suppercarrier (talking about a 1080Ti in my case).

 

I will only try AA, supersampling or PD when the next gen GPUs have arrived.

 

I accept if your opinion differs and I also welcome optimizations to the current software. But I prefer smooth framerates over image quality every day. Much more enjoyable long hour sessions that way for me.

[squirrelmeister]

My Reverb is broken and I wait for the replacement. I use my rift right now while waiting for the Reverb replacement.

 

I am talking about both HMDs:

 

In my opinion the best option to use VR today with currently available hardware:

Get rid of AA, supersampling or pixel density adjustments to improve the image quality.

There are only so much physical pixels on your display and while the rendering at 100% will surly look awful to you at first after a short adaption phase you will stop noticing the shimmering and the jaggies. What do you get in return? Smooth framerates, 60 FPS on the Reverb, 45/90 FPS (motion smoothing) on the Rift, at all times in single player, multiplayer and even on the channel map and on the deck of the suppercarrier (talking about a 1080Ti in my case).

 

I will only try AA, supersampling or PD when the next gen GPUs have arrived.

 

I accept if your opinion differs and I also welcome optimizations to the current software. But I prefer smooth framerates over image quality every day. Much more enjoyable long hour sessions that way for me.

 

 

 

 

It is very interessting how different the frames are for everyone here. Let me explain: I too have a 1080Ti but I am running with MSAA 4x and get constant 40 FPS (due to ASW kicking in) and I know that I am getting 80+ in air. I use the shader mod from Speed-Of-Heat and the shadow modification to just see the cockpit shadows. But I am thinking about getting all shadows back since most of the time I am running at 40FPS anyway and with shadows on high I am still not going under 40FPS - Supercruiser is another story but that depends a lot on other modules being visible (and that is actually generally the case allthough on the SC it seems to be more of a problem than anywhere else).

I am running no extra PD (set to 1.0 in the settings) since the performance impact is way more noticeable for me in comparison to when running with MSAA 4x and the FPS show that too. I for myself came to the conclusion that there isn't any combination of hardware out there to this day that would be able to handle DCS in all its glory at 80+FPS in VR. And because of that the goal should be to achieve at least constant frames over 40 so that at least with ASW or similar technique it is playable and imo very much enjoyable. That is my case and for what I am going for - but there is just one big problem: the ghosting effect is of course very much noticable when high speeds are involved (in contrast to other planes, objects, etc.) but that is the downside of the techniques involved to achieve a good gaming experience with the hardware at hands.

 

 

Another conclusion out of my personal experience and testing the last couple of weeks is that I strongly believe that DCS is way more CPU bound than GPU bound (I hope it is clear what I mean - english isn't my native language). But even on the GPU side there are differences. Especially when you compare the older 1080s/1070s to the newer gen of GPU like 2070s/2080s. I have the feeling that the older ones perform way better than the newer gen of GPUs - no testing here, just what I found reading through this forum and having an eye on peoples rigs. I think this has something to do with DirectX11 being the thing used by DCS and that the older gen is performing better with DirectX11 than the newer generation, but feel free to correct me on that statement - always good to know more and new stuff.

 

 

And another one is: too much third party apps seem to influence your gaming experience in general and not only in DCS. This could be a simple hardware monitoring software which is causing frame drops and stutter because of the monitoring, it could be your OS (and if you don't have for example Win 10 Professional than IMO don't even bother playing games because of the restrictions the OS is giving you), it could be anything, literally.

 

 

Just my two cents on this whole topic.

 

 

And to say it a little shorter: I am just astound by the differences in performances for so many people. But also think that a lot doesn't ultimately mean that DCS is the bad boy in this situations (two examples: HWINFO64 caused massive stutters for me and was the cause for all frame drops I had in DCS; SteamVR standard resolution setting is actually nearly a 1.3-1.5 PD on the Rift S which lead to bad performance from the beginning as you can imagine - I am not using SteamVR anymore because of that).

 

 

:joystick:

 

Cheers:pilotfly:

[Eldur]

I've almost run out of options to disable or degrade, I really don't want to fly in minecraft.

 

I almost got the impression that I can look farther in minecraft than I can do in DCS even at high settings since minecraft doesn't have that thick fog...

 

I also run Tacview and VAICOM Pro but they were only about 4% CPU combined, and I can't do without VCP.

 

Disabling Tacview almost doubled my framerate, just a hint...

 

And for some reason, the Edge mod does hit a lot which is why I experienced utterly subpar performance on 50793 in the first place. Switching to another plane made me have about the same performance as the versions before, dating back to the SC release. The main difference was that that version did run well with Tacview while the newer ones don't.

 

BTW always been running DCS without any AA modes, and always PD 1.0 except for trying out more than that once to see if it improves the Viper HUD which was not worth the performance drop at all.

[Dunx]

Thanks Eldur, I ended up disabling all exports in the end.

 

I too have DCS PD 1.0 and SteamVR at 100% (2212x2164). 60Hz is not too bad in multi-player now, although still a bit of ghosting on busy servers.

 

Textures: High

Terrain Textures: Low

Civ. Traffic: Off

Water: Low

Visib Range: High (not keen on fast growing trees)

Heat Blur: Off

Shadows: Off

Res. of Cockpit Displays: 1024

MSAA: Off

Depth of Field: Off

Lens Effects: None

Motion Blur: Off

SSAA: Off

SSLR: Off

Clutter: 0

Trees: 80%

Preload: 100000

Smoke: 1

Anisotropic Filtering: 8x (not keen on odd looking runways)

Cockpit Global Illumunation: Off

Rain Droplets: Off

Force IPD: 58

Bloom Effect: Off

MSA Mask Size: 0.10

[TOViper]

In my opinion the best option to use VR today with currently available hardware: Get rid of AA, supersampling or pixel density adjustments to improve the image quality. There are only so much physical pixels on your display and while the rendering at 100% will surly look awful to you at first after a short adaption phase you will stop noticing the shimmering and the jaggies. What do you get in return? Smooth framerates, 60 FPS on the Reverb, 45/90 FPS (motion smoothing) on the Rift, at all times in single player, multiplayer and even on the channel map and on the deck of the suppercarrier (talking about a 1080Ti in my case). ... But I prefer smooth framerates over image quality every day. Much more enjoyable long hour sessions that way for me.

 

True for me, I love to have 90fps over Nevada, and don't care about small details. The fluidness of the picture is what I love so much, and it gives me the feeling of beeing there. Sometimes, especially in Nellis and having many static aircraft placed there, the framerate drops significantly.

 

Currently this is true for my config and the selfmade MP mission I fly very often (never on official servers):

= 90fps 90% of time

= 45fps 8% of time (near ground)

< 45fps 2% of time (near statics in Nellis)

 

For details, see https://forums.eagle.ru/showpost.php?p=4384335&postcount=11

[westr]

I've spent all day playing with settings in the OB today to get better performance in VR. Conclusions I've come to if your using Oculus Tray Tool (latest version) Default ASW mode for me is best set to adaptive. Its Between this and OFF but I find with adaptive it is just slightly less juddery. Process Lasso does nothing for me and DCS in fact it often stopped DCS loading. In actual fact I back to back tested Multithreading on and my DCS definitely benefits from it being on. I've managed to push my CPU past the 4.20 mark but the gain has been minimal. My GPU is pretty much maxed out between 90-100 use in game. The only slight improvement I noticed was using alt+return to maximize the game window, not really FPS improvement but certainly helped with stutters.

Ultimately between testing this and the stable version I have a huge performance advantage in the stable version, and my stable version is on the slower disk drive.

Not a dig at ED I know they introduced a lot to DCS World recently, Just really hoping they can find some snags and optimize performance, I know I've got to go to a RTX 2080 at some point but don't want to pay out just yet.

[squirrelmeister]

I've spent all day playing with settings in the OB today to get better performance in VR. Conclusions I've come to if your using Oculus Tray Tool (latest version) Default ASW mode for me is best set to adaptive. Its Between this and OFF but I find with adaptive it is just slightly less juddery. Process Lasso does nothing for me and DCS in fact it often stopped DCS loading. In actual fact I back to back tested Multithreading on and my DCS definitely benefits from it being on. I've managed to push my CPU past the 4.20 mark but the gain has been minimal. My GPU is pretty much maxed out between 90-100 use in game. The only slight improvement I noticed was using alt+return to maximize the game window, not really FPS improvement but certainly helped with stutters.

Ultimately between testing this and the stable version I have a huge performance advantage in the stable version, and my stable version is on the slower disk drive.

Not a dig at ED I know they introduced a lot to DCS World recently, Just really hoping they can find some snags and optimize performance, I know I've got to go to a RTX 2080 at some point but don't want to pay out just yet.

 

 

 

 

Hey there!

 

 

Did a lot of testing yesterday and my conclusion is that DCS is definetely more CPU bound than GPU bound. No matter the setting I use, one of the most important one is "Visible Range"! Everything else seem to have no to very less effect when it comes to FPS. How I got to this conclusion? Just a simple change in PD down to 0.5 gave me just like 10-15 frames more which seems in my eyes to be not normal since every other game will work absolutely flawless if you set the PD to half with almost all settings pushed to the max and I can be certain to reach constant 80 FPS with a PD of 0.5 (looks definetely shitty and not worth playing but for testing I thought this would be a good choice).

 

 

Cheers :joystick:

[imacken]

Hey there!

 

 

Did a lot of testing yesterday and my conclusion is that DCS is definetely more CPU bound than GPU bound. No matter the setting I use, one of the most important one is "Visible Range"! Everything else seem to have no to very less effect when it comes to FPS. How I got to this conclusion? Just a simple change in PD down to 0.5 gave me just like 10-15 frames more which seems in my eyes to be not normal since every other game will work absolutely flawless if you set the PD to half with almost all settings pushed to the max and I can be certain to reach constant 80 FPS with a PD of 0.5 (looks definetely shitty and not worth playing but for testing I thought this would be a good choice).

 

 

Cheers :joystick:

 

GPU or CPU bound depends entirely on your CPU and GPU. Also depends a lot on your settings and the scenario you are flying in. If I fly over a town in the Channel Map, my GPU is up to 98% and gets very hot. CPU is going along quite nicely in the green zone in fpsVR.

Other scenarios, it can be the opposite.

[squirrelmeister]

GPU or CPU bound depends entirely on your CPU and GPU. Also depends a lot on your settings and the scenario you are flying in. If I fly over a town in the Channel Map, my GPU is up to 98% and gets very hot. CPU is going along quite nicely in the green zone in fpsVR.

Other scenarios, it can be the opposite.

 

 

What frames are you getting in vr on the channel map? Especially the instant action free flights?

 

 

 

 

Cheers :joystick:

[Eldur]

Noticed something weird in my first 50979 run. FPS given by DCS were down again at 22 mostly, but it felt a lot better than that. Only in the end, after shutting down my Tomcat the counter actually rose up to 45. Guess it's time to grab fpsvr to get more reliable results