I’ve tried messing around with DLSS, OpenXR TK etc in the past but when spending all my time in the iFly MAX8, TAA is the only way to go for me since that sharpness can’t be beaten IMO.
I also like the idea of not using any fake stuff but instead rely on pure power delivered by the hardware, using the full native resolution my VR headset comes with and so on.
Maybe I will give a 5090 a go when they are in stock to see what if any improvement that would bring vs my 4090.
I get that. But have you tried DLSS 4 and Preset J? DLSS Quality looks darn good. Yes, it’s rescaling 2K to 4K. But the improvement in visual quality compared to DLSS 3 is undeniable. So much so that I feel like I can use DLSS Performance if I need to - something I couldn’t do with DLSS 3, as the quality wasn’t acceptable to me.
I tried it with preset ‘latest’ which should be ‘K’ as I understand it but considering how I didn’t see any real improvement over DLSS 3 when looking at the minimums value on the PFD, maybe I did something wrong 
Probably not.
I’d like to use TAA (or even DLSS DLAA) but as strange as it sounds, my hardware (7950X3D / 3090 Ti) isn’t up to it. For me though, DLSS 4 Quality is the perfect balance of quality and performance. I haven’t heard of the ‘K’ preset. Maybe that’s only available with a newer driver than the one I’m using (566.14)
This is the clip I watched where I heard about the K preset
I also switched from the i9 13900K to an AMD Ryzen. Initially, I had Ryzen 7 7800X3D and on release day, I upgraded to the Ryzen 7 9800X3D. I’m not sure what setup you’re using, but the difference between the i9 13900K and the Ryzen 9 9800X3D at the same MSFS settings in 2D is 30 FPS (avarage). If the difference for you is only 10 FPS, something might be off.
Even when the increase of like 10 FPS was the increase I saw and using all the same settings, you are correct. Reason I didn’t see more is how I had my FPS limited by the refresh rate of my 4K TV which I use for 2D mode and where the refresh rate is 60 Hz.
With the FPS in 2D mode using my 13900K at this point was around 45-50, based on what you say, if it wasn’t for the limited FPS, I would have seen like 75-80 FPS which of course is an amazing uplift!
However in my case and based on how we’re in the VR subforum, what matters to me doing all my flying in VR is VR performance. I wish the uplift would have been the same there but unfortunately, that wasn’t the case.
Well, I’m still using a 5800x3D with a 4090 with a Pimax Crystal. I too have considered upgrading my CPU to a 9800x3D.
Thing is, as I only fly in VR, the demands of the Crystal running at “native” resolution of 4312 x 5102 means I am always very much GPU limited.
Typical CPU/GPU frame-times with DLSS4 transform preset K, using performance mode in FS2024 Beta on medium settings with TLOD of 200, are currently CPU around 13 ms, GPU around 20ms…sometimes a bit higher GPU times sometimes a tad lower, depending on location, aircraft and weather (I don’t fly airliners).
Upgrading the CPU seems unlikely to do much to help the GPU. Maybe I could raise some settings, like TLOD, which more heavily impact CPU performance than the GPU, but I think it would be minimal improvement to FPS in VR.
Currently I can now get high 40s in rural areas in a C172 and over 50 in a glider over the Alps, which given how good it looks now on the Beta with DLSS4, means I don’t really need to upgrade either CPU or especially to the stupidly high expense of a 5090.
I believe the improvement differences in VR may be a lot less (typically about 1/2). Do you have a VR headset yet?
I was (and still am) GPU limited too, but performance and especially smoothness increased a LOT going from 7800x3D to 9800x3D for me, despite being locked at 45fps both before and after.
More than worth it! Something like 20-25% reduction in CPU frametimes, all microstutters gone (unless I hit the GPU limit too hard for an instance and drop to, eg, 44fps for a second), probably due to the improved lows, which are super important in VR. Everything feels much more stable and solid with more consistent frame pacing.
Don’t make the mistake of thinking a better CPU can’t help just because you’re GPU limited. I can’t recommend this chip enough for VR 4090 users.
I hear you and am encouraged by your experience, BUT, was this in FS 2024 or FS 2020? You returned FS 2024, so we may be talking apples and oranges.
The one thing I noticed about FS 2024 compared to FS 2020 was the increased smoothness, despite the 30% or so drop in performance (plus the bugs of course). This was born out in Cap Frame X, which showed better 1% and 0.1% lows, and no red stutter area in the pie chart.
So upgrading to a 9800x3D may indeed benefit me in FS2024, but not so much as in FS2020? I’d be interested to see what other 9800x3d owners have to say using it in FS2024 VR, with a 4090 and a Pimax Crystal.
Either way, I’m not closing my mind to the CPU upgrade, despite needing a new mobo, and RAM. It is after all a lot cheaper than a 5090. 
No this is all 2020 only, I did not return to 2024 (yet)!
This summer is a planning to build my new game pc. Because the 5080 suxs, the 4090 is not on stock anymore its gonna be a 5090 or maybe a new 5080ti. 5080 with 16gb vram is almost an insult🤣
I will make a giant leap for sure.
It’s all theoretical calculation.
My last upgrade was 200% better in theory, but in practice +/- 30 fps.
I hope Asobo and Nvidia fix the sims and drivers so everybody can enjoy flying with or without lots of money spending on hardware. 
I benchmarked my system and it was verry close to the calculated values. The gap between and 5090 or 4090 and a 3080TI is huge. Together with a slighly faster cpu and much more faster ram and ssd i think these calculated values are in line with the real thing
So are you saying you have a 5090 you’ve been able to compare to a 4090?
The theoretical gap is large, but practically speaking the performance is not that far apart.
I think that the 5090 is not much better then the 4090.
It’s all about a few fps and a bulk of money.
It seems to me that 5090 might be a better choice for VR.
For 2D and MSFS 2024/2020 RTX 4090 is already sufficient. So, if someone has an RTX 4090 and doesn’t play VR, upgrading to the RTX 5090 doesn’t seem to make much sense.
But of course, it’s better to wait for proper tests, without fake frames 
Testing GPU or CPU in a game that is full of bugs, especially issues with VRAM, makes absolutely no sense. I would understand a test in MSFS 2020, but I don’t get the point of testing MSFS 2024. Also, I don’t think there’s a comparison between 5090 and 4090 – or maybe there is a FPS comparison that I missed?