VR - experimental support

User avatar
KingOfTheRoad77
Posts: 232
Joined: 07 May 2021 13:27
Contact:

Re: Oculus Rift - experimental support

#1941 Post by KingOfTheRoad77 » 12 May 2022 19:16

@Krists :
I'd be interested to see your fpsVR stats of a transport from Omak to Everett, and from Everett to Olympia, taking en route just the exit for Tacoma and directly back to the freeway, too, during daytime, if possible. Those are the two trips I did with the 5800X 3D. The area around Seattle is quite crowded and has a lot of traffic, so I am curious of the GPU and CPU frametimes you'll get there with the Reverb.
Krists
Posts: 35
Joined: 30 Nov 2021 02:14

Re: Oculus Rift - experimental support

#1942 Post by Krists » 12 May 2022 21:28

Just did these two trips:
Yakima to Seattle to pick up special delivery at night and rain and then actual delivery from Seattle to Everett:
fpsVR Report:
App: amtrucks HMD: HP Reverb Virtual Reality Headset G20 (90.001 Hz, IPD 68.0)
GPU: NVIDIA GeForce RTX 3080 Ti (30.0.15.1259, Tavg 71.4, Tmax 76) CPU: AMD Ryzen 7 5800X3D 8-Core Processor (Tavg 59.8, Tmax 65)
Delivered fps: 57.01 Duration: 28.3min. Headset was active: 100%
GPU Frametimes:
Median: 11.5 ms
99th percentile: 18.6 ms
99.9th percentile: 20.1 ms
frametime <11.1ms(vsync): 46.7%
CPU frametime:
Median: 9.4 ms
99th percentile: 21.9 ms
99.9th percentile: 25.3 ms
frametime <11.1ms(vsync): 63.9%
Reprojection Ratio: 2.5% (for Index/Vive/VivePro headsets only)
Dropped frames: 8 or 0.0% (for Index/Vive/VivePro headsets only)
Max. SteamVR SS: 150%
Render resolution per eye: 3876x3792(by SteamVR settings, Max.) (HMD driver recommended: 3164x3095)
And then during sunny morning time back from Everett to Olympia with normal dry van trailer:
fpsVR Report:
App: amtrucks HMD: HP Reverb Virtual Reality Headset G20 (90.001 Hz, IPD 68.0)
GPU: NVIDIA GeForce RTX 3080 Ti (30.0.15.1259, Tavg 73.0, Tmax 76) CPU: AMD Ryzen 7 5800X3D 8-Core Processor (Tavg 60.7, Tmax 65)
Delivered fps: 79.08 Duration: 12.7min. Headset was active: 100%
GPU Frametimes:
Median: 9.7 ms
99th percentile: 12.7 ms
99.9th percentile: 13.7 ms
frametime <11.1ms(vsync): 91.2%
CPU frametime:
Median: 7.3 ms
99th percentile: 11 ms
99.9th percentile: 13.8 ms
frametime <11.1ms(vsync): 99.1%
Reprojection Ratio: 0.2% (for Index/Vive/VivePro headsets only)
Dropped frames: 1 or 0.0% (for Index/Vive/VivePro headsets only)
Max. SteamVR SS: 150%
Render resolution per eye: 3876x3792(by SteamVR settings, Max.) (HMD driver recommended: 3164x3095)
Honestly I didnt feel any difference between either of trips - perhaps slow special delivery and rain somehow "muted" the frame drops, but it did feel comfortable.
Nevertheless compared to my previous PC setup - 3080 FE, 2x16GB 3200 CL14 and 5600X, I can feel a big difference, however I am too lazy to swap the parts back to do actual testing with fpsVR.
I think the biggest difference must come from GPU update.

EDIT:

One more trip during sunny day time from Omak to Everett with fruit juice delivery:
fpsVR Report:
App: amtrucks HMD: HP Reverb Virtual Reality Headset G20 (90.001 Hz, IPD 68.0)
GPU: NVIDIA GeForce RTX 3080 Ti (30.0.15.1259, Tavg 71.8, Tmax 77) CPU: AMD Ryzen 7 5800X3D 8-Core Processor (Tavg 60.0, Tmax 65)
Delivered fps: 67.52 Duration: 26.6min. Headset was active: 99%
GPU Frametimes:
Median: 9.7 ms
99th percentile: 14.9 ms
99.9th percentile: 16.1 ms
frametime <11.1ms(vsync): 84%
CPU frametime:
Median: 5.8 ms
99th percentile: 12.7 ms
99.9th percentile: 15.4 ms
frametime <11.1ms(vsync): 95.9%
Reprojection Ratio: 0.4% (for Index/Vive/VivePro headsets only)
Dropped frames: 12 or 0.0% (for Index/Vive/VivePro headsets only)
Max. SteamVR SS: 150%
Render resolution per eye: 3876x3792(by SteamVR settings, Max.) (HMD driver recommended: 3164x3095)
And in ETS2 one more trip from Salzburg to Verona with Renault T and with metal tubes load. Started in morning during sunny weather and ended in evening during rain and below you can see the result.
I am a bit surprised on why there is such big difference, but I think I have somewhere maintained higher graphics settings for ETS2 as visually game is much more anti-aliaised and looks much better than ATS - especially Austria looks really stunning now.
However I don´t understand what settings I have selected or changed to eliminate the jittering/aliasing and would like to increase ETS2 performance by slightly reducing graphics and increase graphics quality in ATS to eliminate aliasing.
This is especially visible with wires, fences and tree shadows in forestry roads.
fpsVR Report:
App: eurotrucks2 HMD: HP Reverb Virtual Reality Headset G20 (90.001 Hz, IPD 68.0)
GPU: NVIDIA GeForce RTX 3080 Ti (30.0.15.1259, Tavg 73.9, Tmax 77) CPU: AMD Ryzen 7 5800X3D 8-Core Processor (Tavg 60.6, Tmax 66)
Delivered fps: 47.89 Duration: 26.6min. Headset was active: 100%
GPU Frametimes:
Median: 17.5 ms
99th percentile: 21.7 ms
99.9th percentile: 23.9 ms
frametime <11.1ms(vsync): 10.8%
CPU frametime:
Median: 8.7 ms
99th percentile: 13.8 ms
99.9th percentile: 17.2 ms
frametime <11.1ms(vsync): 89.9%
Reprojection Ratio: 30.6% (for Index/Vive/VivePro headsets only)
Dropped frames: 17 or 0.0% (for Index/Vive/VivePro headsets only)
Max. SteamVR SS: 150%
Render resolution per eye: 3876x3792(by SteamVR settings, Max.) (HMD driver recommended: 3164x3095)
User avatar
KingOfTheRoad77
Posts: 232
Joined: 07 May 2021 13:27
Contact:

Re: Oculus Rift - experimental support

#1943 Post by KingOfTheRoad77 » 13 May 2022 06:45

Hi :)

Thx a lot for posting your results 👍

Could you do me one last favor?
=> Could you raise your SS% further so you get around 4514 x 3948 = ~17,8M pixels? And do another trip around Everett/Seattle/Olympia or southwards towards Portland? That would fit my 8K-X's resolution and would give me another, better hint whether it's really my GPU that is limiting or whether the 5800X 3D still might give me some benefit and it perhaps just was bad luck those two days.
I've got the feeling that your effective resolution with 150% SS is close to the tipping point, or perhaps even a bit beyond. Especially as you mentioned that the upgrade from the 3080 FE to the 3080 TI did give you another FPS improvement - of which I think it wouldn't help me neither with my HMD... I'll probably need a 4080/4090 🥴

Thx a lot once again for your support :)
Krists
Posts: 35
Joined: 30 Nov 2021 02:14

Re: Oculus Rift - experimental support

#1944 Post by Krists » 13 May 2022 08:21

I will try later today ATS with higher SS, but I can already see that I’m GPU bound in both ETS2 and ATS. Also if you look at reprojection rates (I believe they are the main driver for stutering and unpleasant VR experience), for some reason I have huge issues with ETS2, although the game engine, graphics settings in config and for VR are the same, but visually ETS2 looks much better and much more anti-aliased. I wouldnt even need higher settings for ETS at the moment.

Can you tell what kind of ram do you have?
User avatar
KingOfTheRoad77
Posts: 232
Joined: 07 May 2021 13:27
Contact:

Re: Oculus Rift - experimental support

#1945 Post by KingOfTheRoad77 » 13 May 2022 09:11

2x 16GB DDR4 3600 MHz Corsair Vengeance RGB RT

Yes, from my personal estimation, I also have the feeling that ETS2 has a little worse performance than ATS. But there, too, it largely depends on the country and area I'm driving in: The older maps generally work much smoother in terms of FPS and frametimes than the newer ones, Spain in desert and steppe areas is better than France or the forest parts in Western Russia, and with Grimes Winter mod I get an avg. FPS of 60-65 whereas it drops to ~50-55 with his Spring and Summer mods. So there are a lot of influencing factors.

Besides: All ingame graphics options maxed, MSAA "medium" - especially concerning MSAA, I cannot notice any difference between MSAA disabled, low and medium... and even "high" is almost negligible. But I admit, too, that I am a real "sharp high quality image junkee": Especially when I sometimes get constant 70-75 FPS for a couple of seconds in UK, Nevada or Arizona, the general low-resolution blurriness of many textures is quite obvious. So, driving with (much) lower graphic settings certainly improves FPS and frametimes - but in contrast to this reduces the overall image quality and blurriness much too much for my personal taste. It's a bit as jumping from the frying pan into the fire - while I prefer mediocre, but still enjoyable and playable FPS over mediocre image quality as long as I have to choose and cannot have maxed both at the same time for the moment.
Krists
Posts: 35
Joined: 30 Nov 2021 02:14

Re: Oculus Rift - experimental support

#1946 Post by Krists » 13 May 2022 09:25

What is your CAS latency for ram?

The more I think about image quality, the more baffled I am on how different they are in each game with the same settings (I posted config of both games in previous page). However I tried to take through the lens photo and video with my phone and these are results (pay attention to railings of the structure in ETS2 and road line markings and wind deflector in ATS):

https://imgur.com/a/0q1qyGX

Although completely dofferent locations, you can see how smooth the shadows and other lines are in ETS2 vs ATS and it seems like besides Steam VR supersampling, I have additional antialiasing applied for ETS2, however for both games the stereo buffer is set to 1.0 in config. I also checked Nvidia settings and in both games antialiasing is disabled (“disabled” status in ATS and “-“ status in ETS2).

@Komat - are there some hidden antialiasing settings/differences between each game? I thought that the engines are identical and they should perform more or less the same. I would like to increase performance in ETS2 and increase graphics quality in ATS, but for both games I have the same Steam VR global settings (150% SS) and stereo buffer at 1.0.
Last edited by Krists on 13 May 2022 12:14, edited 3 times in total.
User avatar
KingOfTheRoad77
Posts: 232
Joined: 07 May 2021 13:27
Contact:

Re: Oculus Rift - experimental support

#1947 Post by KingOfTheRoad77 » 13 May 2022 09:29

Krists wrote: 13 May 2022 09:25 What is your CAS latency for ram?
Manufacturer's stock values for XMP2: 18-22-22-42
Krists
Posts: 35
Joined: 30 Nov 2021 02:14

Re: Oculus Rift - experimental support

#1948 Post by Krists » 13 May 2022 09:39

I have g.skill trident z 2x16gb 3600mhz CL14-15-15-35, perhaps this also gives us some “boost”, to 5800x3d, but I doubt it would be very significant.
User avatar
KingOfTheRoad77
Posts: 232
Joined: 07 May 2021 13:27
Contact:

Re: Oculus Rift - experimental support

#1949 Post by KingOfTheRoad77 » 13 May 2022 09:55

After all I've read through different reviews, 5800X 3D doesn't benefit as much from faster RAM than "older" CPU's due to its larger Cache. So I'd agree that it certainly has some influence, but probably not a major one.

EDIT:
I've found a comparison in a German review, comparing different RAM speeds between 5800X and 5800X 3D:

[ external image ]

=> Whereas one might notice it going from 2300 or 2666 MHz up to 3200, 3600 or 3800 MHz with a 5800X (somewhat between +10 and 16%), with the 5800X 3D the difference tends to be mainly "only" measurable (far less than +10%). At least I probably wouldn't really - perhaps more a slight "feeling" than a hard fact. And I doubt it behaves much different when it comes to the CAS latencies within the same MHz.
Krists
Posts: 35
Joined: 30 Nov 2021 02:14

Re: Oculus Rift - experimental support

#1950 Post by Krists » 13 May 2022 10:23

Also another interesting thing is that 3080TI performance doesnt really hit 100% and is jumping between 75-95% (although I had 11gb of memory consumption in ETS2 and 9.5GB in ATS) and CPU is even less 10-30% if I remember correctly, which is a sign of how optimized the game is for VR.
Post Reply

Return to “General discussion about the game”

Who is online

Users browsing this forum: NonanteNeuf, Volvo1990 and 11 guests