An information that maybe will save some money for someone: in VR the performance of 2080 ti is worse than those of 1080ti. and not only , in Ac, Ir and Rf2 the performances are almost identical if not worse. I think ill go back to my "old" 1080ti.
Well all the performance gain statistics are shown for dx11 and dx12, they probably don’t put any effort into checking how dx9 works on new cards/drivers. Shame it didn’t gain you anything
While you are right, this thread is about VR, which is vastly different from 2D in raceroom. CPU is actually even more of a bottleneck in VR, to a point where it is just impossible to get 90 fps on high settings with a full grid with current hardware afaik. Especially settings like shadows absolutely tank fps in VR, luckily you can still make some use of excess GPU power by increasing SS values
Or an engine upgrade Luckily 90 fps in any situation is possible, though you do have to sacrifice a lot of settings, most of which you honestly don't notice quickly in VR anyway. I run with mostly minimal settings (shadows completely off apart from contact shadows), track detail medium and textures high, 2x AA and 200% SS and get 90 fps in VR everywhere with my 1080ti + 4770k @ 4.4GHz. So anyone with an equivalent or better CPU will be able to do this too, you forget about graphics as soon as you actually start driving in VR anyway edit: RAM speed is also surprisingly important in VR, i upgraded from 1600mhz to 2400mhz and gained 10-20% faster frametimes!
I’m just gonna add my findings - going from a titan (pascal) to a 2080ti. I have a i7 6950X overclocked to ~4.2ghz. The titan and 1080ti are pretty close in performance (the titan however has a lot of overclocking headroom in my experience) . I never/refuse to do 45fps, so I will sacrifice visual for proper 90fps. So all my tests in VR are based on 90fps In Raceroom I went from 1.8 super sampling and 4xAA to 2.0 super sampling and 8xAA and I turned track detail from medium to high, track animation on, increased visible cars up to 20 (from 17). I now have everything on high except shadows and reflections which are set to low. I can actually run +2.1 except for Sonoma, the Chinese track (the straight) and down hill post turn 1 at Spa with a lot of cars in front of me. I am overclocking my 2080ti but not as seriously as I did my titan. I’m seeing right around 30% increase in speed. My benchmark -
Have you tried OC'g the 6950X further? Though it's a 10-core, I understand R3E only runs 1 (or 2?), so perhaps raising the clock for 1 or 2 cores might show some increase in performance?
That is insane if you can run those settings at 90fps :O I wonder if running the rift in steamVR causes some really bad overhead or something, because my 4770k really isn't that far behind your 6950X in terms of single core speed.
I have - went to 4.5ghz but not only do I see no difference in Raceroom but I’d get stutters in single player. At 4.3ghz that seems to go away but even water cooled the machine gets noisy as the radiator fans really get going. My benchmarks go higher but it’s all cpu side.
I see. Perhaps that explains why my stock 8700's usage barely reaches 80% (while GPU usage just slightly above). BTW, I'm using your old settings. So I haven't really found any need to OC. Really hope for some optimization.
This the question I was thinking of. Because to play R3E in VR with the Rift we have to use SteamVR which is an additional layer (Oculus Home + SteamVR) and Nathan is using a Vive so in my opinion it is more efficient... No way to bypass SteamVR?
I have a vive pro and only run steamVR. I don’t have asw on. I don’t really understand how Raceroom single core is working. If I run with single core the game goes to absolute crap. Here is a direct compare of our cpu’s https://cpu.userbenchmark.com/Compare/Intel-Core-i7-4770-vs-Intel-Core-i7-6950X/1978vs3604 I assume your right when it comes to single core. However, what I see when looking at my resource manger is 4-5 cores pressed about 70-90% and 4-5 others doing 20-30%. I assume the ones doing very little are systems and crew chief etc. But sure looks like more than 2 cores are being throttled by Raceroom. I would sure like clarity -
The single-core performance is actually almost identical, as you compared to the non-overclockable 4770, while I have my 4770k at 4.4GHz. You also have 3400mhz RAM vs my 2400MHz, That could make a significant difference. I think the reason why you see multiple cores being used is that windows just shuffles a single core application around all cores to spread the load. Though maybe RaceRoom does use multiple cores for certain stuff, should ask sometimes
SteamVR is the only option unfortunately, apparently you could prevent oculus home from opening while playing steamVR games before to gain performance, but with the oculus dash update they prevented that, so now it's impossible afaik
It is possible to both avoid Steam VR (I am working on that, by adding Raceroom support to Open Composite, but just that you have realistic expectations, that won't yield higher FPS, that should give smoother framerate) and avoid Oculus Client. To avoid Oculus Client, current inconvenient way is: Option1 * Make Oculus Client.exe run as admin * Start R3E in VR mode, that will start SVR. * Run Oculus Dash .exe when R3E starts. (rogramFilesOculus\Support\oculus-platform-runtime\oculus-platform-runtime.exe"') Option2 (so that you don't have to wait for R3E start): http://www.mediafire.com/file/lpblil1sqq59685/OculusRuntime.zip/file Two more hacks: * Oculus Homeless: https://www.reddit.com/r/oculus/comments/8uf1sm/oculushomeless_use_dash_without_home_20/ * Revert to Home 1.0 (I wouldn't recommend this one however). https://i.reddit.com/r/oculus/comments/9odbvz/fixed_old_oculus_home_runtime_w_instructions/
No. Some people claim Home 1.0 is better, but I have not noticed any perf reduction with 2.0. The fact that I really dislike Oculus' outrageous push of ads is another matter.
By the way is there any benefit to increase the oculus server priority? Oculus tray tool allows you to do it.