First a couple of examples from my own experience (with HP Reverb G2 and 2070 Super)

1) When AA set to TAA (in VR) I set renderscale to 50 because I see no difference between 100 and 50. Is that not strange not seeing any difference between so different values ?

2) When I enable NVIDIA DLSS I am GPU limited while still in 2D but when I switch to VR I get main thread limited (has never happened before). Should it not be the other way around ?

Benchmark results

3) MSFS benchmarks shows 5800X3D superior to all other CPUs. But according to
https://cpu.userbenchmark.com/Compare/Intel-Core-i5-12600K-vs-AMD-Ryzen-7-5800X3D/4120vsm1817839
even a 12600K is more powerful than a 5800X3D. Yes that benchmark is about other things not MSFS. But is it not strange that there are such big difference ?
In the linked benchmark commentary they warn us to believe all of the praise of 5800X3D

1) This is indeed strange. I see a marked difference if I set renderscale <~80 with my Reverb 1. Even 90 has a certain impact. I never tried 50, but 60 is already really bad with a lot of shimmering. 

I am usually Main Thread limited on my 3440x1440 screen but GPU limited under VR, as I would expect.

My system is in the sig. No explanation for your results... In any case you should see a markedly higher fps (+50% or so) with Renderscale 50.

Kind regards, Michael

42 minutes ago, jfri said:

When AA set to TAA (in VR) I set renderscale to 50 because I see no difference between 100 and 50. Is that not strange not seeing any difference between so different values ?

VR often has too many resolution/renderscale settings. Some in the SteamVR settings (both global and per game), some in the OpenXR settings (both in WMR and OpenXR Toolkit), and in the case of MSFS there's also one in the game. So it could be that your VR resolution is rather high as a result of some of the settings outside of MSFS, so it doesn't make much visual difference if the MSFS setting is at half.

34 minutes ago, JacquesBrel said:

VR often has too many resolution/renderscale settings. Some in the SteamVR settings (both global and per game), some in the OpenXR settings (both in WMR and OpenXR Toolkit), and in the case of MSFS there's also one in the game. So it could be that your VR resolution is rather high as a result of some of the settings outside of MSFS, so it doesn't make much visual difference if the MSFS setting is at half.

But Steam VR is not used with MSFS. There is little visual difference between 50 and 100 but the impact on performance is very noticeable. Which hint at resolution scale in MSFS indeed change the VR resolution used

1 hour ago, jfri said:

3) MSFS benchmarks shows 5800X3D superior to all other CPUs. But according to
https://cpu.userbenchmark.com/Compare/Intel-Core-i5-12600K-vs-AMD-Ryzen-7-5800X3D/4120vsm1817839
even a 12600K is more powerful than a 5800X3D. Yes that benchmark is about other things not MSFS. But is it not strange that there are such big difference ?
In the linked benchmark commentary they warn us to believe all of the praise of 5800X3D

It's not strange.  Flight simulators are not like other apps or games that they typically benchmark.  We are interested in sim performance not benchmarks, though, right?

Yes the 5800X3D has been discussed here a lot by people I trust.  As far as I know (and that's not much, as I don't own the cpu and haven't really researched it fully), the primary reason why the 5800X3D performs so well in the sim is its large L3 cache, which is as I recall 96MB in size and that happens to be 11.5x more L3 cache than my 8700K has and more than even the new Intel 13th-gen CPU's have.   For some reason, as I understand things, the MSFS architecture thrives with this big L3 cache.  There's a  thread here on this topic called "AMD 5800X3D -- Brilliant Performance" which you should look at.  The good people here at AVSIM in that thread are not bought-and-paid-for (at least I don't think so!) and should give you honest information.

 

6 hours ago, jfri said:

3) MSFS benchmarks shows 5800X3D superior to all other CPUs. But according to
https://cpu.userbenchmark.com/Compare/Intel-Core-i5-12600K-vs-AMD-Ryzen-7-5800X3D/4120vsm1817839
even a 12600K is more powerful than a 5800X3D. Yes that benchmark is about other things not MSFS. But is it not strange that there are such big difference ?
In the linked benchmark commentary they warn us to believe all of the praise of 5800X3D

The Intel bias of that platform is well documented. E.g. here Intel Core i5-10600 sample manages higher bench result on UserBenchmark than AMD Ryzen 5 3600 despite overall lower test scores - NotebookCheck.net News

7 hours ago, jfri said:

But Steam VR is not used with MSFS.

Yes it is, if you have a Steam VR headset.

5 hours ago, JacquesBrel said:

Yes it is, if you have a Steam VR headset.

I don't need and do not start Steam VR with MSFS. In P3D XP11 and DCS I must start up it otherwise it launches by itself.

I have a Reverb 1 and you can switch back and forth between MS OpenXR and Steam VR OpenXR. I tried both and they both work but settled with MS OpenXR. SteamVR adds yet another layer of complexity to the already complex VR environment, notably making switching into/out of VR more error-prone.

And yes, P3D requires SteamVR.

Kind regards, Michael