Qcumber

Sorry to hear that. The 9070xt looks like a good GPU. Maybe it just needs a slightly different approach to settings compared to Nvidia!? I'm happy to help problem solved if I can.

4080 to 5080 maybe not. But 4070 to 5080 was more than double performance for less than double the price.

Yes. A single place with links to useful threads at the top. Updated regularly.

Yes. My main aim is to tease out the settings which impact most, irrespective of rig. It's not an easy task. Ideally it would involve using preset values and test tracks that people could run on their rigs and compare results.

That's a good point regarding CPU and GPU usage but I'm not sure how I can capture this alongside the other data. Any suggestions about software I could use? It would need to allow data export to a CSV file. It would be interesting to correlate usage with render times. My observation is that if GPU is maxed out then the render time exceeds 13.88ms and fps drops; which is exactly what you would expect. CPU usage would've harder to correlate due to multiple votes.

On the suggestion of a Discord member (Orion_42) I have tried this test again without using QVFR and by changing some of the settings. The charts below are 3245x3302 pixels set using OTT. The blue trace (T2) is with visibility range set to Ultra and all the scenery slider set to max. The red trace (T3) with visibility range set to Ultra and all the scenery slider set to lower settings as per the screenshots below. There is only a marginal difference and there is much less stutter and random frametimes than I saw in the previous post using QVFR. To be sure, I retested QVFR using the same settings as above. smoothen_focus_view_edges=0.1 sharpen_focus_view=0.9 horizontal_focus_section=0.4 vertical_focus_section=0.4 peripheral_multiplier=0.5 focus_multiplier=1.1 The dark purple line is with max visibility settings. The pale purple line is minimum visibility settings. The black dotted line is the "blue" curve from above for reference. Overall these results are much better than yesterday and more stable. Interestingly QVFR with these settings has very little benefit in terms of GPU render time (all three situations were around 13 ms. CPU render times were 8-9 ms. Some stutter but still comfortable.

With DLSS 4 you need to push the resolution high to minimize ghosting. If you use QVFR in VR you can maximize the foveated region to optimize performance. Make sure you use preset K. I had a 4070 until recently and you can get a decent performance with this GPU.

OK. The F-4 take off is painful. Its worse than a take off on the Marianas map! What the hell is going on. It does not seem to be very well optimised in terms of objects etc. Green is F-4 free flight. Blue is F-4 take off (Afghanistan map instant actions). Not all of the GPU render times are being captured so there is something odd here. On the ground both CPU and GPU render times are all over the place. No consistency at all. Once airborne and high enough FPS stabilises. The free flight is a dream performance. Solid FPS and low render times.

I'll give it a go.

Yes. You can get away with most settings at high. I'll try to post my settings later. I did some relative performance testing a few weeks ago with my 5070ti. These were quite basic so I am hoping to repeat these in a more structured way soon with the 5080.

I use head tracking combined with JoyToKey to reset and offset cursor position, with a wireless "finger mouse" strapped to my joystick. I've been using this setup for ages and it works well. I can reach any switch in the cockpit. I also map a lot of common tasks to the HOTAS and "double up" by mapping one button to be the shift key.

I'm not sure how much difference it makes but that is what I do. You only really need 1 TB as you OS/boot disk. I 4 M2 PCIE slots on my motherboard so use a 500Gb for OS, a 2TB for DCS and a 1TB for other games etc. Given how big DCS is getting, and how many maps you intend to buy, 4TB is more future proof but is expensive. I have room for another 2-3 maps before I need to upgrade.

Eye tracking can be used for Dynamic Foveated Rendering using software called Quad Views Foveated Rendering. This allows high resolution in the centre region and lower resolution in the periphery. It can give you a big improvement in performance. Some people prefer not to use it as you can sometimes notice the low Res in your peripheral vision but this can be tweaked with settings. Yes but I don't find it very useful. Others have though.

Is this the PTC version? I have been avoiding using PTC but might give it a go if it's working OK with the QP.

+1 for Quest Pro.

+1 for the Quest Pro. It remains a very overlooked headset but is one of the best. Particularly with regards eye tracking. It makes a massive difference in terms of performance. Forget pixel counting and comparing paper specs. Real world performance with this headset is excellent. I managed to get mine used for £300. As new condition. You might get lucky on eBay.

I ended up getting a 5080 and sending the 5070ti back. It's a good card but I managed to get a 5080 FE for oy £150 more than the 5070ti. The 5070ti works very well in VR. I use a Quest Pro with eye tracking. It can manage 72 fps in most situations.

I have also seen big dips in performance when flying towards large bodies of water inland. e.g. flying NW from Berlin towards Warren there is a large group of lakes (sorry but I am not sure what they are called). Very strange. It's the only performance issue I have seen so far on this map.

What do the CPU render times look like when this happens? With a 9800x3d they should be around 6-8 ms (my experience with the F4). Are the GPU render times OK? At 60 fps locked they should be below 16.6 ms.

What about a 4070? 12Gb RAM and pretty good performance in VR. I used to have one before upgrading.

I am also getting very good performance with the 5080. Overclocking potential is very good. I have a +10 oc which is stable in DCS. The last NVIDIA driver update has also given me a +5% boost in stock and OC performance. In theory this puts it very close to a stock 4090. I previously had a 4070 and the 5080 is +130% faster. I briefly had a 5070ti and the 5080 was about +30% faster in both synthetic benchmarks and in DCS. I can manage 90 FPS in many situations with plenty of GPU headroom. But in heavy situations, such as low over a big city, performance drops. 72Hz handles this a lot better. I can manage 72 fps pretty much all the time with very few dropped frames. VRAM use is very good and I have not seen this go above 12Gb. If anyone is interested I have done some testing with the 5080 and am in the process of looking at the effects of various DCS settings on performance.

Sorry if you have already tried this, and it might not relate to your issues, but I have had boot issues with 6000 RAM and EXPO. Boot times would vary sometimes taking several minutes. After enabling "memory context restore" in the BIOS boot times are now about 30s.

I can also confirm that VRAM use on the 5080 is very good. I have not seen it go over 12 Gb either. At launch the 5080 was about 12% behind the 4090. The latest NVIDIA driver (276) has given me a 5% boost in stock performance. Overclocked I get a 10% boost which translates to almost 17% increase relative to stock speed using the previous driver. In theory this puts the 5080 very close to the 4090 stock performance. Of course this is with synthetic benchmarks. In DCS (VR) the overclock is comparable but I have not tested if the latest driver update shows any meaningful change.

The F4 is very demanding in VR. What you describe has happened to me a few times. I am not quite sure what causes it but it is similar to what has been reported when entering and exiting the map view for some people. This has been attributed to VRAM overflow but I have not noticed any excessive VRAM use. What are your DCS settings and headset resolution? Do you use QVFR? Does it happen on all maps? I would suggest 2 things if you have not already tried them. 1) submit your DCS log file to the Discord log checker. 2) reduce the headset resolution and DCS settings to see if it still happens.

Here are my current settings. These are not the maximum I can get in synthetic benchmarks but are stable in DCS.