Yeah you should. Even with my sub optimal setup I was getting almost constant 60 (a few dips to 50’s) in New York flying very close to everything, with TLOD and OLOD 200 in the Cowan 206 helicopter.
So many settings to change so maybe hopefully it’s just finding that sweet spot.
Not VR though I’ll add!!! 
Are both monitors and the VR headset plugged into the graphics card?
Or is any of those attached to the motherboard DP/hdmi?
The reason I am asking is to determine whether or not the iGPU is somehow interfering with your experience.
yes, all 3 plugged directly into the 4090. One of the settings I found in the bios this afternoon was to disable the integrated gpu, but didn’t notice any change in perf after doing that. I’ve been using cpuid and earlier I could see the integrated Radeon gpu listed along side my 4090 but after disabling it, it no longer appears in cpuid.
I also seem to notice when i check the cpu load that it often doesn’t seem like it is working very hard even when I’m in the sim in vr. It can peak at 100% but is low most of the time.
OK Thank you.
Maybe you’ve already done this, but I would try to run the sim with the high resolution monitor only, measure performance, then plug the HDMI monitor, measure performance, then plug the VR headset, and see how it goes.
In order to determine if the performance is vastly different, or not in each of these cases.
Actually before I send you on a wild goose chase, I just tried again with my current settings (DX11, no frame gen) and I am about the same as you.
35 or so in the densest part when looking at everything, usually around 40, up to about 50 or a bit higher if looking away from detail.
I am sorry to mislead but I guess I have tried so many things I forgot what I had set up before - when I say 60 before that was WITH frame gen frames included (my first out of the box test - this is first place I tried 2 weeks ago and it blew me away! 
)
However I will say even now with DX11, it is very fluid - no skipping/hitching, no tearing, even at 36fps it’s fine for me - totally flyable and no stutters.
I’ll give that a try tomorrow after work and report the results.
I’m currently using dx12, I’ll try switching back to dx11 tomorrow. I do think that will probably be a little better, it was before on my old pc.
It is definitely flyable, but it sure would be nice to have more frames, I envisioned flying around NYC getting 80 or so frames with all the settings maxed out.
yes, using in built frame counter, and also using the openxr toolkit just for its frame counter that is vr friendly.
Gosh, must be pretty disheartening to spend that much $$ only to get the same sorta fps you can from much cheaper systems. This is the problem with buying brand new tech, it tends to be a bit unstable. Oh how I have bad memories of 1st revision motherboards for brand new chipsets.
Hopefully this will get sorted for you with driver updates.
This why I also stay away from AMD. No doubt it works for most people most of the time, but whenever I’ve had AMD systems I’ve always had to deal with weird hiccups. In fact, right now some dude in one of my flightsim groups just got a brand new AMD system and it having similar issues. When it works though, the X3D chips do seem to fly.
I don’t think you need 80
My goal is a stable and most importantly consistent 60. Anything more you hardly feel.
I have been running locked 50% refresh of my 120-Hz set monitor and that’s what I am trying to dial in.
It’s still gonna be a struggle but this is with everything Ultra except the highest Terrain Shadows (which I don’t think really affects it, so might bump that up too).
Just tested unleashing to 100% refresh in an area that isn’t dense but still has some nice visuals (the new Kern Valley freeware) and it can reach crazy 116fps (~105 on the ground), but really the feeling between that and 60 is not worth it IMO). More likely to notice the fluctuations if you really try to push that.
For me it’s not disappointing - this is a HUGE upgrade from what I was using before and even with these teething problems it’s frustrating but I know we will sort it soon. I could never dream of getting this smoothness AND with the details maxxed out before so I am not complaining, it’s just hard to know if it’s performing as it should.
Fair enough. That system should fly once the issues are worked out.
I’ve been having the time of my life in this sim even on my old machine and this new machine is giving me way better visual quality in VR and at the moment flyable frame rates in heavy areas, just had very high expectations that haven’t fully been met yet.
I’m going to keep researching and experimenting, and keeping an eye out for the next bios or chipset update, or the next suggestion on this thread. I’m optimistic that more can be coaxed out of this machine.
hmm, this is casting some doubt on the perf of 7950x3d at higher resolutions
Back again with another update.
I’ve uninstalled process lasso and the parking tool to get back to baseline and focus on bios settings.
I decided to connect only my 38 inch monitor and try to maximize perf on it first.
The results today are very strange to me, using CPUID to watch, it looked like neither my GPU or CPU are working very hard yet my frame rates are low. It was fairly smooth at these low frame rates and visuals were fantastic.
Settings were ultra with terrain LOD at 400 and object LOD at 200, and it seemed my frames stayed about the same no matter if I used TAA, DLSS Quality, or DLSS Ultra Performance. Even switching from DX12 to DX11 didn’t change things much.
TAA DX12
TAA DX11
DLSS Quality DX12
DLSS Quality DX11
DLSS Ultra Perf DX12
It is a good idea to focus on one monitor.
If you can make your system work in a satisfactory fashion with the large screen in DX11/TAA, the rest will hopefully fall in line.
It is intriguing that your Cinebench scores, while not the highest I’ve seen for this combo, are not aligned with the performance in the MSFS screenshots, even though the CPU scheduling is skewed towards CCX#0, where the VCache resides, as it should.
Have you tried to run another game, or a benchmark like TimeSpy?
Just to cover the basics, and determine if the issue is only with the MSFS install?
I’m not really a gamer, don’t have any others installed. I’ll give TimeSpy a try and report the results.
I just ran another cinebench since I had changed some bios settings.
During the multi thread test it had all cores pegged to 100%, but during the single core test, the one core that was working some was no where near 100%, less than 50% most of the time while the test ran.
Multi core test 34278
Single core test 1823
For reference, here is a Cinebench run done by OC3D.net:
Yours is not far off, so it would be worthwhile to see how the system behaves in another application.
The base program for TimeSpy is like 35$, not really wanting to spend that much to benchmark.
Check out this link. Last time i checked it came with a free demo, but right know I do not have a Windows computer in front of me to try it myself.
oh, they do have a demo in steam, getting it now.
Edit installed it but it is taking forever to “Initialize”
