G2 and 1080Ti

[Dragon1-1]

I've recently got the Reverb G2, and I'm currently in process of trying to optimize the settings for the best experience with it. However, I ran into a problem: all recent guides assume I have a recent GPU, such as the 3090, and I don't. I have a 1080Ti, which is a bit weaker. Still, the G2 came out before the 3090 did, so I'd expect it could be made to work with the second best GPU of its time.

 

Right now it's playable, but barely. I'm running about 20-30 FPS (as indicated by the counter you get with RCTRL+Scroll lock). I'd like to nudge it up to 45, but neither lowering the resolution nor fiddling with shadows, MSAA or anisotropic filtering seems to have had any effect. Even disabling SSLR didn't budge it. I've tested on "Home on the Range" (NTTR mission for the Viper) and on a mostly empty mission in Syria, on the ramp at Ramat David.

 

My question is, does anyone have good suggestions at what settings to try in order to increase the FPS? My CPU is i7-4770K, and I haven't messed around with overclocking yet.

[dburne]

All you can do is lower graphics settings in game and G2 resolution.

You will want to be able to maintain 45 fps with motion smoothing on.

[speed-of-heat]

my old settings were for the 1080ti, on 2.5.6 and the newer ones are for the 3090 .. i hope they will help, if you have K  level processor, you really should overclock it if you have a good cooling solution, otherwise you are leaving money on the table

Edited June 13, 2021 by speed-of-heat

[largecod]

Hi
I have the combination.
G2 and 1080ti.
I totally doubled my performance frame rate from around 20fps to on average 40-45fps.
I achieved this by upgrading my cpu from a amd fx9590 to a Intel 10900k overclocked to 5.1ghz.
I bought cpu and mobo from scan.co.uk which came with water cooler and 32gb ram.
They overclocked on site and bench tested for 24 hrs.
For £1200 this is the only way forward considering the pandemic we are in.
I'm not paying £3000 for gpu.
No way.
And tbh dcs is old technology and cpu upgrade will help not gpu.
My settings are near enough maxed out apart from super sampling in steam VR settings.
They are set to 70%.
If I get time today I will send over my settings.
But tbh cpu upgrade best option.
Admittedly on the supercarrier looking at the tower structure it drops to 30fps and heavy Syria missions


Sent from my ANE-LX1 using Tapatalk

[eaglecash867]

I can't say that I noticed a significant difference in performance between my previous 1080ti and my current 2080ti.  There did seem to be an improvement in depth perception, but that was about it.  That was all on my old CV1, but once I got the Reverb G1 and got it tweaked right, I got really good performance with it.  

 

The G2 really isn't much different in terms of performance.  Per a suggestion I found in another thread, yesterday I set my SteamVR Supersampling to the automatic default (which on my rig was 128%), and then set DCS pixel density to 0.6.  That yields excellent clarity and performance...even improved spotting of ground targets.  I'm going to try tweaking things up a little higher to see if there is any benefit, since it looks like I have some headroom to play around with.

 

I was going to take the plunge into a 3090, but unless you were in the right place at the right time, the scalpers got most of those.  The rest of the supply problem is probably another "...because of COVID..." thing.  

[Dragon1-1]

OK, so I'll try the PD trick, and I'll also have a look at overclocking the CPU. I have a small water cooler, so I've been meaning to get around to squeezing some more performance out of the CPU. IIRC, this particular example had a problem of not overclocking very well, but it was air-cooled back then. I don't think I can get 5GHz, but maybe I can get it to 4-something.

 

Upgrading the CPU outright isn't on the table ATM, mostly because I'd need a different mobo to go with it (also for DDR4 memory, I'm currently stuck with DDR3). I won't say I have no budget for upgrades, but for now, I'd like to get the most of what I currently have, even if I do have to leave off some eyecandy. The next upgrade will probably mean a full-on rebuild of my PC, except for the GPU and the "slow storage" HDD. I'm not quite ready to pull the trigger on that.

[speed-of-heat]

let me be frank the PD "trick" just that, it has been shown many times to be just that.  Leave PD at 1.0 and reduce the Custom resolution Percentage in steam, you can do it the other way around, but, the maths is easier if you use steam's Custom Resolution Percentage

 but either way just change one of them 

[AlpineGTA]

Agreed. Leave PD at 1.0 and adjust the per app resolution percentage in Steam.

 

For example, I only have a AMD 5700XT but I can achieve 45fps if I set the per app Steam resolution to 50%. This might sound low but it's still surprisingly excellent clarity (I can read all the gauges and displays without leaning in) and is still better than the (still very good) Quest 2 clarity. The only difference between this and running the Steam res. at higher percentage is you will notice a lot more jaggies so I run MSA x2, but 50% Steam res. gives me enough overhead to be able to do this. I can run Steam res. at 100% and MSA off but my fps is about halved.

 

Everyone's system is obviously different but I would say the basic rule is keep PD at 1.0 and juggle the Steam per app res. until you get something you're happy with.

[eaglecash867]

49 minutes ago, AlpineGTA said:

Everyone's system is obviously different but I would say the basic rule is keep PD at 1.0 and juggle the Steam per app res. until you get something you're happy with.

I lived by that rule since the first day I got my Reverb G1.  Tried the other way just the other day and I like how it looks and feels better.  In any case, no matter which way you choose to manipulate render resolution, it represents a really small time investment and can easily be reversed/changed if you don't like the results.  There is also zero risk of damaging hardware or causing system instability like overclocking does when you're just getting started on tweaking things for best performance.

[AlpineGTA]

I think it's good to try and give a starting base to work from, but yes, I think that's fair enough, go from there to whatever works for you at the end of the day.

[speed-of-heat]

To each their own; providing you understand that .5 in PD is not equivalent to 50% in steam then all is good ... which leads to the problem if you mess with both you can very easily get confused as to what resolution you are actually operating at ... but as i say to each their own.

[eaglecash867]

Not really concerned with what resolution I'm running at.  I'm shooting for a balance of maximum clarity and performance and I'm doing that by "feel".  This is all after a couple of years of living by the "set DCS PD to 1.0 and leave it alone, adjust SteamVR SS up and down" mantra.  Never even thought it was possible to go lower than 1.0 PD in DCS, until I read in another thread about Vive representatives talking about why SteamVR defaults to a specific SS percentage to correct for barrel distortion on the new VP2.  So, I tried the suggestions in that thread for my HMD, setting SteamVR SS to its default value for the Reverb G2.  So, I ended up going from 70% SteamVR SS, with DCS PD of 1.0...to 126% Steam VR SS, with DCS PD lowered to 0.6.  Not sure how to explain it, but it just looked better, and ground targets were easier to spot from greater distances.  I was able to bump the PD up to 0.7 and still get really good performance and clarity.  It'll be interesting to see what my other VR titles look like now that I'm not dumbing them all down to make DCS happy.  I know that SteamVR allows you to do "per app" settings, but that just seems like a lot of extra work when its just one title that is struggling in VR.  I'd rather make the one straggler in the group compensate for its inability to keep up with the others...just makes more sense to me.

[Supmua]

52 minutes ago, eaglecash867 said:

Not really concerned with what resolution I'm running at.  I'm shooting for a balance of maximum clarity and performance and I'm doing that by "feel".  This is all after a couple of years of living by the "set DCS PD to 1.0 and leave it alone, adjust SteamVR SS up and down" mantra.  Never even thought it was possible to go lower than 1.0 PD in DCS, until I read in another thread about Vive representatives talking about why SteamVR defaults to a specific SS percentage to correct for barrel distortion on the new VP2.  So, I tried the suggestions in that thread for my HMD, setting SteamVR SS to its default value for the Reverb G2.  So, I ended up going from 70% SteamVR SS, with DCS PD of 1.0...to 126% Steam VR SS, with DCS PD lowered to 0.6.  Not sure how to explain it, but it just looked better, and ground targets were easier to spot from greater distances.  I was able to bump the PD up to 0.7 and still get really good performance and clarity.  It'll be interesting to see what my other VR titles look like now that I'm not dumbing them all down to make DCS happy.  I know that SteamVR allows you to do "per app" settings, but that just seems like a lot of extra work when its just one title that is struggling in VR.  I'd rather make the one straggler in the group compensate for its inability to keep up with the others...just makes more sense to me.

 

I'm running SS 150%, per-app 100%, IPD 0.7 on my Vive Pro 2 and the cockpit texts are tact sharp with next to no shimmering.  At least it is better for me on the Vive Pro 2, and not just in DCS same with MSFS also.  Now I just leave SteamVR SS alone and use in-game setting to reduce res instead if that option available.

[eaglecash867]

19 hours ago, Supmua said:

 

I'm running SS 150%, per-app 100%, IPD 0.7 on my Vive Pro 2 and the cockpit texts are tact sharp with next to no shimmering.  At least it is better for me on the Vive Pro 2, and not just in DCS same with MSFS also.  Now I just leave SteamVR SS alone and use in-game setting to reduce res instead if that option available.

Yup.  That's what I'll be doing too.  It just makes more sense to reduce resolution in DCS when its the only thing that struggles.  I also like the added benefit of even better depth perception and target spotting ability since trying what you had suggested in the other thread.

[Dragon1-1]

What's the best scenario to test VR performance? I just found "Home on the Range" doesn't seem to be it, as no matter what I do, it gives me about 20 FPS on the ground, more if I'm not looking at Vegas. At one point I got 45 FPS when I took off, but then, I fiddled around with the settings, tried another mission, and it was much worse, despite no performance change on the ground in my "benchmark".

 

I want to arrive at a setting where a "typical" SP mission with an average number of units will play smoothly. I need some scenario that would translate well to how it'll be in regular usage.

[speed-of-heat]

i would suggest you do this:

 

1) fly a simple mission, say for 3-5 minutes

 

2)  make a track from that mission

 

3) use fpsvr to benchmark that mission at a known set of settings with motion smoothing/vectoring/ whatever turned off, and the headset stabilised and send the results to a log file (you watch the 2d monitor) 

 

4)  change one setting, re-run the mission and review the average difference in frametime for CPU and GPU from the baseline if any variance is less than .3ms re run the mission for that setting three time and take the average

 

5) rinse and repeat 4

 

see how i do it here for the broad idea:

 

My 3090 Settings for my G2 (2.7.1) - Virtual Reality - ED Forums (eagle.ru)

 

remember you aren't really interested per say in the FPS but you want your GPU to maintain less than 22ms and CPU less than 11ms, and try and create as much headroom as possible HTH

 

 

 

 

 

whilst adding units seems like the sensible thing it creates way to many variables 

[Dragon1-1]

I use FPS as a proxy because that's what DCS shows. FPSVR is payware, I have better things to spend my money on (although if there's a freeware version, I suppose I could try that). I'm not really interested in chasing numbers, anyway, what I want is something that feels right and that would translate to complex missions in something like Raven One. This is something I'm afraid a simple mission won't catch, as it would task GPU more than CPU, while a trigger and AI heavy campaign mission would put, proportionately speaking, more load on CPU.

[speed-of-heat]

NMCNMM

[zildac]

I use FPS as a proxy because that's what DCS shows. FPSVR is payware, I have better things to spend my money on (although if there's a freeware version, I suppose I could try that). I'm not really interested in chasing numbers, anyway, what I want is something that feels right and that would translate to complex missions in something like Raven One. This is something I'm afraid a simple mission won't catch, as it would task GPU more than CPU, while a trigger and AI heavy campaign mission would put, proportionately speaking, more load on CPU.

It's an invaluable tool, especially if you're having issues.