Gunnars Driver

We should have about the same computer performance. If I lower my CPU clock to 5.0, it shouldnt be much diff. I have a 2080ti Strix OC, that further clocked but not very big differences. Im way lower than your frame times, and higher in FPS: A bit busy, the winter just arriwed. Ill try to make a flight in a default Huey instant action, for comparing. There seem to be a big hickup for your system that needs to be found. Did you clear the shaders ? (I have never done it, not needed it but I saw people with problems do it here and that fixed most of the issues).

Threes visibility is hard on fps, and so are grass/cliutter. Try them on minimum. Set MSAA to off also, for this test, and SS to 100%. Chimney smoke to the lowest possible. cockpit global illumination to off. I didnt se any change when I tried MSAA mask after it got implemented. I think its supposed to mask the part of the picture you dont see in the goggles, and thereby reduce the demanded CPU/GPU calculations. Dont use it for these tests anyway.

Ehh, you shouldnt go lasso before it work as it should, even if you could gain from lasso( dont know, havent tried) you should make sure everything is ok before. Can you Print screen the dcs settings? Is your 7700K setup to do 5.0 on all cores all the time? Use some kind of software to check this, perhaps HW monitor 64 or something. The CPU should preferebly do 5.0 on all cores and all the time. What cputemps do you get ? Check that the CPU isnt working with other strange stuff when you do nothing = CPU load should be close to zero when looking in this forum. Check this post about co,puter setup. Might kick in open doors but... https://forums.eagle.ru/showpost.php?p=4498999&postcount=20 Try DCS settings like these. Do not use the presets down to the right. Arrows to the left, ,imi,u, setting and arrow right, try max on preload radius. Set SS in steam to 100%. Run a instant action mission, like take off or landing or something and then check framtimes etc with fpsvr

Have you mixed with any CPU affinity in DCS ? If so, that might cause this ? You shouldnt get fram times that high if any core isnt close to 100%. If its not 100%, then its not working all the time, and with frame times over 11ms(if 90fps headset), it should really do that. Someting wrong nut at this point hard to see what...

It might be easier to see the load core for core woth windows task manager, performance tab, resource monitor, and open the cpu tab. Widen the window, rund DCS for a while, and then alt tab to the resource monitor and look for the CPU loads. This is a very old post from me, but look at the picture: https://forums.eagle.ru/showpost.php?p=4163309&postcount=288

Seems very strange to have the CPU load well spread amongst the cores...or it would ve noce if it had bern this way. We know the main thread of DCS only can use one thread, so you should really have one core loaded. What RAM do you have, size/clock speed? SS200% is quite high. What about using more regular settings and do a test on a instant action, to see that everything works like it should. Refering to servers that produce a big load or high frame times without knowing everything works perfect when these loads isnt present might make us hunt ghosts. Low load on the core driving the main DCS thread should’nt be consinstent with frame times above 20ms It could be nice to have your hardware down in your signature, so we dont meed to go back to the initial post to find out. Also, what VR goggles do you use?

Of course this is badly handled by DCS but as long as it doesnt affect the gaming ”in flight” I dont mind. I do not think I have seen VRAM usage be higher after I switched aircraft and is flying it?

I wouldnt be surprised if we find the ryzen 5900x or so not to produce as good single core values as iin the advertisment/tests, this due to DCS not using the most modern instructiuons. We can sure hope it does it in DCS also....would be very nice but... Anyway, even if ot does, its 4.8Ghz in single core and 4.5 in 12 core so if you get that one and it delivers as it is said wont be that much faster than your 7700K at 5.0. The expensive part is the GPU, a descent PSU isnt that much and a Noctua D15 is 90 bucks. You dont need to get a 3080, a 2080ti will be fine if you like to save money. I have a 2080ti now and it runs very nice with my reverb( 150% SS) I have the 3090 on order, but also a pimax 8KX which I recon need all the GPU power there is. A 3080 would be nicer of course in your case. Also nothing wrong with a 5900x but it calls for a new mobo and I guess new RAM so virtually a new computer build. For DCS purposes youll be fine with your 7700K clocked to 5ghz.

So, we imagine you have made the cooling sufficient. The 7700K isnt known to run cold. Use good cooling paste(liquid metal like coolaboration liquid pro or conductonaut is best) It could be wise to have a good temperature monitoring software. I like Core Temp: its free, and it show temps for each CPU core in the windows bar to the right beside the clock. https://www.alcpu.com/CoreTemp/ https://www.ocbase.com/ I checked this video(manual overclocking video) and it looks good: https://www.tweaktown.com/guides/8075/ultimate-msi-z270-gaming-m7-motherboard-build-guide/index.html Follow it step by step. I would begin to set the AWX stepping to -2 or -4 to make sure AWX instructions dont give you problems. I dont think DCS use AWX, so you can have them set lower to not hang the CPU in test or when doing other stuff then DCS, and still get 5.0Ghz or so with DCS. You can set a higher voltage then 1.300 if the CPU hangs on 1.300v. More than 1.400V start to be quite much, so staty below that number. Also, increasing voltage increases temps and I guess you might run into temp problems( lets say 90 degrees is safe, and 100 is the limit). The overclocking needs tests to ensure it runs stable. Real ocerclockers runs the overclock fdor days with very though stressing software but for DCS that at most only loads ione core much, you should use for example occt ( https://www.ocbase.com/ ). Use this for the stress tests and use the setting not to stress test with AWX to begin with. If it hangs, increase the voltage by 0.02V or so, and run the test again. Keep an eye on the core temp values for the CPU, and abort if it runs to hot. Then you need to lover the voltage or get better cooling, ultimate way is to delidd the CPU...but it isnt reallly for forst time overclockers if you ask me( not hard to do but a slight risk of bricking the CPU). I had to delidd my former CPU (i7 3770K) because it did run very hot and the delidd saved me around 25 degree C at load. 7700K is also known to run hot due to Intel using the same bad way of transfering heat from the cpu chip to the IHS. That a very short course on OC. I havent worked much with MSI motherboards, just a little. I more or less always do ASUS formyself. Should be good for a start and if the cooling is up to it and you didnt get the worst 7700K you should be able to hit 5.0.

Your is a kind of budget cooler. Not bad for the size but it could be wise to get better cooling. Remember, for DCS purpose (single core performance) a 7700K is more or less as good as any CPU , especially if you can get it to Ghz. If you have the space and would like to go for a simple solution, the Noctua D15 is the way to go. Very good cooling and low noise. I got if for my older i7 and kept it when I got the i9 9900KS, overclocked to 5.2Ghz. Still cooling good and low noise levels. There are other good aircoolers but the D15 is in the top and possibly the quietest. My friend builded a i9 10980XE 18 core and bought the same cooler, very efficient. He also got the NZXT Kraken 360mm Z73, which was very noisy but cooled a only a few degrees better than the Noctua D15. After changing the fans to noctua, its not noisy anymore but the corst for a NZXT Kraken Z73 +3 140 mm Noctua isnt a cheap solution at all. I’ll dig into the motherboard and overclock later.

Good. I’ll check that up and be back.

I started overclocking in the 90’s. A good Overclock takes a little time. You should start with making sure your cooling is up to it. What cooler do you use? There is no standard settings for overclocking. First, different motherboards wendors have different names for the same thing, and between different chipset it differs despite the same vendor. Its not that hard though. Most 7700K can do 5.0Ghz. You need to increase the voltage, and for this there some different ways to do this. Theres also a voltage compensation you should set. What motherboard do you have, full name? We might be able to link to a guide for your mobo. What cooling ?

I see many here complaining about visual detection range. I began with oculus rift and while first gen consumer VR set sometimes made it hard to spot things far away I experience the HP reverb actually give me longer detection range than I have in real life flying. I think most people overestimate by far the actual distance where you can see and detect another aircraft. A normal range would be 5 miles, maybe 10. This in very clear air that give you unlimited visibility( we often have this far north were I live). Rest of the world most often give a visibility thats only a fraction of this). In many places a really nice day gives perhaps 10 to 20miles visibility maximum. You can not see another aircraft beyond the visibility, in fact its hard to detect one thats half that range. With the reverb I see aircrafts quite far away.

While having used cinebench and other measuring softwares, I havent read or learned the whole story about what instructions it use. When comparing new hardware its common to take in account new fancy instructions that are more effective and give more performance per clock cycle. But, for this rating to be ”true” the software or game used need to use these instructions to be as effective. If in fact using software that do not use those fance instructions, the advertised gain in computing power isnt there. I havent checked any facts but I have a slight feeling that DCS lacks the ability to use a lot of the modern instructions. If this is true the gain advertised wont be. My guess is that clock speed still is paramount to DCS.

[Edit]Reference to my old setup, i7@4.8 and GTX1080)[End Edit] Shadows Flat should be possible. Resolution of cockpit displays: I use 1024 every frame( I dont know if there is a difference but I have thought so.) Terrain textures seems to not make a big difference in VR but also not a very big load either. I had high set with my former i7@4.8ghz still with buttersmootj gaming. Threes visibility is very low for you, after the gpu upgrade increase to at least 60-70%, to avoid se them grow :-/ I have preload radius maxed out and this have helped with the smooth gaming for me. It might load the GPU to much( dont know what gpu you have now). After upgrade= do it. Du you have fpsvr ? (small steam software, about 4usd). Use it for determing the CPU and GPU frame times. Change settings and see how the frame time changes. For example, at 90Fps you cannot go above 11.1 ms, and for 45fps 22.2. 1000/fps= max possible frame time. If you dont have or cannot ise with index, use task manager -> system monitor and look at the CPU load om the side that shows load om each core/thread. Look at the single core that is most loaded. When it reaches 100% that would be about the same thing as havibg the frame time getting close to or above the possible according to above.

If its possible to increase the OC a bit maybe 5.0Ghz, its virtually no difference between your CPU and *any* other new when playing DCS. If you domt do other games, there really is no need or only a slight, very slight difference if you get a CPU that go beyond 5.0Ghz. I have a 9900KS, clocked to 5.2Ghz, and Those 4% can be seen in measured data when tweaking but I think if I notice it during gaming it is mostly that I want to believe I notice it. ( that said, CPU clock speed is currently the only weapon that can be used when the CPU is limiting DCS. You can get a win in other games and other software with a new computer build but for DCS I would have tweaked the overclock as far possible. If needed get a better cooling solution for the CPU. After that, put the rest of the money on the graphics card.

Yes. Not to learn you how it looks but its essential for the question if the settings actually make a difference you can see or not. I probably missed that part about Reverb in your post. For this thread, nvidia say they have a mor efficient use of the VRAM in 3080 compared to the 20X0 generation so 10Gb is said to not be less than 11Gb in 2080. Im not convinced so I went for 3090.

I'd say as a beginner the normal FOV of 100-115 is enough. I did dream of greater fov since the first day of VR(2016), but still I have enjoyed these normal FOV VR sets very much. There isnt really many choises as almost every VR set hase more or less the same FOV, I think Valve Index is the the one sticking out with 130 degrees from the normal ones. ' Otherwise, Pimax with around 200 degrees(194 measured?) is perhaps the future. I will get my Pimax in about two weeks after a long wait. I have read a lot during the time and some cant get the settings to produce a sharp image. I think there might be quite much settings, and tuning on the head and also, pimax software wasnt as easy to use as for example oculus. My recomendation is to begin with a easy setup - easy use and start from that point. Also, as you are in the DCS community - DCS needs quite much of your computer so if you go Pimax( more pixels and resolution due to the wide FOV) you probably need the best computer you can by. I think I would recommend Reverb G2 as a good VR set for DCS - you need quite high resolution to see the instruments, gauges etc and also to detect and classify other aircrafts. I have the Reverb "G1" and Im quite happy with the resolution and sharpness(sweet spot a bit narrow in the lenses but that is said to be fixed on G2)

No need to change the motherboard for a 3090(or 3080). You will be fine.

There’s a quite big difference between the first gen VR sets and the latest when it comes to what settings that actually makes a graphic difference. Some settings make a big difference and you only lower them if needed, but also put a high strain on the CPU/GPU and increase the RAM usage. Visibility range = you like as high as possible, but it bite the HW. Threes visibility, to see them grow in front of you isnt fun so you like that setting maxed out if possible. There are some settings where I dont clearly see any difference in the graphics so these I try to not max out. For me with a 2080ti 11Gb Im VRAM limited in some cases. I have lowered some settings to not get the allocsted VRAM to reach 11Gb. I have seen stutter without the CPU/GPU at 100% but VRAM 10.9 or so, and after changing settings that got the reported VRAM use lower the stutter dissapeared. While I cant be sure that is was (only) VRAM shortage that caused this, it was enough for med to not select a 3080 due to the 10Gb only reality and Ill het my 3090 next week.

The IPD adjust in DCS is more of a size adjuster( never touched it, only from what I read). All IPD settings should be done in the ”VR Settings” and IPD in DCS to adjust if things seems to big or to small. As you seems to get the eye convergence as it should be, I guess the issue is that the objects rendered on ground isnt recognizable to the brain when it comes to the size. I also think FOV as you say can be a limiting factor for the depth perception. Even if objects in peripheral vision is not seen sharp they give you a depth perception if they are closer or further away. For real flights, you loose those items (Around 40 degree FOV om each eye) so the binocular feeling is there. You get much more dependent on the indicated airspeed with NVG than without. I will get my pimax 8KX in a about two weeks and at least I think the greater FOV will increase the high speed feeling at low level in DCS( that isnt really there in DCS as it feels IRL) or in racing games. I not sure but the increased FOV might increase the height perception as well.

I’m mostly above 10Gb, 10,3-10,7 (out of 11). I did use 125% SS before and went up to 150%. I dont think its SS alone, but other Dcs graphic settings affect as well.

I had tracking problems recently and that trick helped me. A lot. I had no changes in the rooms lighting or furnishing. I had quite much lights, always on when gaming. I did not use room boundaries, seated gaming only. Tracking got worse and worse and I got the tip to reset room boundaries. This fixed the issue!

If you can judge the height to make smooth landings you at least have some clues about how high you are from the ground. I think the problem that relieves the height dizzines might come from a few issues: The ground is in most cases textured so you can not distinguish any parts and recognize what it is, and the size of them. The brain judges hight and distance with both known size compared to the size on the retina and also the eyes converging angle (at distanses up to around 5m. It work worse but can have some impact up to 10m). If you cant know the size of the objects, that part of the brains height judgement is lost. The converging eye part for sure works very well for me at close distances. If I lean forward towards the instrument panel I get the feeling that Im really close to it. I both guess and feel that this is quite good simulated by the VR systems and the games. Full flight simulators are not as good as VR is on this part. The picture is presented on a dome with a fixed distance, and regardless of what is shown the eye convergence is the same. IRL flying in the mountains often makes the distance judgement very off, because you cannot se if a rock is 1 by 1 meter or 30 by 30m. With NVG the eye converging part is not there and the close distance judgement need some training to work safe( close formation flights, for example). There is a demo for oculus rift(Dreamdeck?) where you are standing on a sort of bridge on a roof of a high building. That part at least gave me a good feeling of heights. I think itight work to get the feeling of height by putting out a refueling truck on the apron in DCS and then taxi close to it. The size of the truck should be easy for the brain to understand. @Taz1004: Did you try the supercarrier with crew on deck ?

Fpsvr ( ~ 4 usd on Steam) is very much worth the money. Otherwise, use task manager —> resources monitor and look at the CPU so you can se the load on all cpu cores. Only one will be highly loaded. You can also see the most basic parts of the gpu there. GPU-Z Is otherwise a good software to use to see more details about the GPU.