Vertigo72

Your cpu "running at 60%" doesnt really mean anything. You have a 4 core/4 thread cpu. DCS can at most fully utilize 2 cores. And only one of those will be performance critical and probably close to 100% all the time. Windows will shuffle those threads across the cores for various reasons, including spreading the heat, so looking at task manager you will see 4 cores that seem underutilized, but that doesnt mean you are not cpu bottlenecked. The 100% utilization rate on the GPU and low FPS does strike me as odd, depending what those "custom high" settings are. If you are going to combine high MSAA with 4K resolution, well, I dont know if even a 2080 can pull that off with the F14. My 1070 certainly struggles even at 1080p, and enabling mirrors on a carrier creates a nice slide show. Here is my suggestion: use settings that others can easily replicate (say DCS default high), chose a mission or training that others can also load, and then we can see if your performance is out of line or not. Since few of us use a 4K display, perhaps note your FPS on 1080p as well.

Save yourself some money and use opentrack with a webcam and some 3 printed LED clip. If you dont have access to a 3d printer or lack any soldering skills, you can buy them on ebay. You can also buy kits that include a camera with a IR filter mod. Usually a modified PS Eye camera, because its high framerate and low price. The entire kit is usually around $35 I have used both a trackir and opentrack with a modified ps eye, and they work pretty much identical. If anything, opentrack has more features.

Two physical or logical cores? Two logical cores isnt going to be enough, and its going to have dramatic effects if those share the same physical core, you would essentially be testing a single core cpu. 2 physical / 4 logical could be enough, although I wouldnt be surprised if there where some hickups. 3 physical / 3 logical (no hyperhtreading) is almost certainly enough to be indistinguishable from 6/12 or more.

Not really. Even on heavily threaded workloads like rendering there is really no point in overriding the kernel. Manually setting core affinity is such a blunt tool compared to the smarts a kernel scheduler and dispatcher can have. The only case where that makes sense is if you know something the kernel does not; as used to be the case when windows didnt properly handle logical cores sharing execution units on a single physical core, or when the OS isnt aware of NUMA in some CPUs like threadripper. Another scenario where it *might* make sense if your desired outcome differs from process priority. Say you are doing a 10 hour 3d rendering job and gaming at the same time. You probably care more about the game frames being rendered asap than the 3d render. But then really, the better solution there is not setting core affinity, but setting process priority.

DCS only uses a single main thread that uses ~60-70% of the cpu time that DCS uses. It spawns a few other threads, including one for the GPU driver, a C runtime library and one for sound. Those combined account for ~30-40% at most, the bulk of it the GPU driver. The rest is essentially zero, literally a few microseconds of CPU time after a few hours of playing. So basically DCS can only benefit from 2 hardware threads. One that is performance critical and runs the main DCS.exe thread, and one that will be idle much of the time doing the rest. If you want to be generous, add a third hardware thread for windows background tasks and kernel housekeeping, although that can likely run on the second core with no serious performance penalty and you pretty much reached the limit of DCS parallelism. So you dont need hyperthreading for DCS, you dont need 6 cores/12 threads. A 4 core / 4 thread cpu is going to be exactly as fast (everything else being equal) and I even doubt it would be significantly faster than an otherwise identical dual core, if you could still buy one. Dont believe me? Try it. Start|Run|msconfig|Boot|Advanced options|Number of processors Set it to 4 or even 3, and see if you notice a difference. Pretty sure you wont. Set it to 2, and there maybe a small difference. All that said, as I said earlier, there is absolutely no reason to assume it makes performance worse if the OS is aware of hyperthreading so Im not advocating disabling it. I just wouldnt want people to go out and buy a threadripper or Skylake-X because "DCS uses 70 threads"

It is completely unnecessary for process scheduling / core affinity when you use windows 10 and any regular desktop CPU thats properly supported (edge cases might include NUMA processors like threadripper). There may be situations where the "gaming mode" helps, by disabling clock scaling. Particularly on laptops. But this would only make a difference if you have a buggy bios or a bios that doesnt support your CPU, and the better solution there is upgrading your bios. If thats not an option, you could achieve the same by disabling speedstep in the bios (for intel) or cool and quiet I think (AMD), but you shouldnt have to.

reserved

I want to know how fast your PCs are using a standardized test. Since no one responded, I made my own, short track, which is pretty much worst case scenario using only free modules. Its also quite representative of my flying skills :) This may be more a CPU than GPU test, as my GPU utilization rate is pretty low; this could be due to the amount of planes, objects and missiles, but this isnt uncommon in complex missions or online, so that doesnt make the results any less relevant. If you want to post your results, please follow these rules: - if possible, disable all overclocks and run your hardware at stock speed (easier to compare) - Set DCS graphics settings to HIGH preset (vsync disabled). - Change your resolution to 1920x1080 - I strongly recommend you restart DCS after making any changes and delete fxo and Metashader2 folders or the settings may not be applied. - Double check your drivers that you are not forcing vsync or other AA settings through the GPU drivers. - Use Afterburner and start the benchmark when the track starts to play, stop it immediately at the end. Use the log file it creates. If you dont know how to do that, this will help: Then report your results along with your hardware config. Below is my result. Feel free to use it as template DCS 2.5 open beta 2.5.4.30038 Ryzen5 2600X stock speed, water cooled. 3.6Ghz base clock, 4.2GHz boost, actual clock ~4.1 GHz 16 GB DDR4-2133 Geforce 1070 stock speed M2 NVME SSD 01-05-2019, 17:13:35 DCS.exe benchmark completed, 12898 frames rendered in 219.532 s Average framerate : 58.7 FPS Minimum framerate : 12.4 FPS Maximum framerate : 112.9 FPS 1% low framerate : 11.5 FPS 0.1% low framerate : 5.3 FPS Reran this test after receiving my DDR4-3200 10-05-2019, 16:47:19 DCS.exe benchmark completed, 14646 frames rendered in 218.609 s Average framerate : 66.9 FPS Minimum framerate : 13.9 FPS Maximum framerate : 136.1 FPS 1% low framerate : 13.1 FPS 0.1% low framerate : 10.4 FPS You may want to run it a few times, and post your best results. Benchmark short.trk

Is there a way to use the scroll wheel to zoom in while you are in free look mode? Right now, I use my mouse to click things, and the scroll wheel to select a fixed zoom level (I also use my throttle's paddle axis to momentarily zoom in, but thats not useful when trying to click). I can only zoom in while Im in one mode, and click things in the other mode, which is irritating because I often need to zoom in before I can click something. So I need to constantly switch modes. Is there an easier way?

You cant use the oculus quest for PC games. Im interested in the rift S, if/when the price drops, as a cheap entry in to VR, but would like to try it first, as I have some concerns about the software only IPD adjustment. Pimax is probably a better purchase, and I was close to buying that, but I dont want to overspend on a VR set yet, as I think its moving too quickly. The is showing me what I really want, I just want it at a price I can afford. And I want integrated eye tracking and foveated rendering to keep GPU requirements in check. I wouldnt mind spending >1000 Euro on a set that I will hang on to, but if I buy something now, I want it to be cheap enough that I dont regret it 12 months later when its made obsolete. We are almost there... But not quite yet.

There is no doubt a 1070 Ti will do better than a 1070. The question is, how much. My speculation is that it will be more significant than most other games, because of the significantly higher texel fill rate (which doesnt do much for most AAA games), and a 1070 Ti should be pretty close to a 1080. Admittedly, the result will still be difficult to interpret, as the Ti not only adds TMUs and ROPs, but also more shaders. My theory is that shader performance is not a major concern for DCS, but thats kinda hard to prove when all cards with higher texel fillrate also have higher shader count or performance.. Oh well. I guess that means it doesnt really matter if you shop for the wrong metric, if the right metric is included anyway ;).

Im not sold on ultrawide form factor. Its just a very cheap way to advertise enormous diagonal sizes for monitors that really are not that big. I have three 16:9 monitors myself, primarily for productivity. But in flight sims, this setup doesnt really work. The way my monitors are positioned, similar to a curved monitor, I dont really get to see to my sides, I just get a wide aspect ratio, similar to the one monitor you linked just a bit more extreme. What that means is that with zoom levels that make sense (not seeing wingtips on 45 degree angles in front of me), I lack downward visibility to my instruments and upward visibility when dogfighting. To solve that, I would actually need to buy one or ideally three more monitors and position them above my current monitors. But guess what, when you have 6x 16:9s arranged like that, I might as well have a single much larger 21:9 tv/monitor and not have to suffer the bezels! Extreme aspect ratio achieved through multiple monitors IMO only makes sense if you can arrange them close enough and angled enough so they surround you. Then you can zoom out and have your wingtips rendered on your 9 and 3 oclock, instead of 11 and 1. Extreme aspect ratio on a single monitor (that can not curve around you *), doesnt make sense for flight sims period. Its like you would have a 55" TV and then cardboard over the upper and lower parts of the TV because somehow, thats better? Reminds me of Top Gear where Clarkson taped over the windshield of his "battle truck" to make it look cooler: * even if the monitor did curve around you, then you would run in to severe projection distortion issues when rendering a single viewport. Games are rendered to be projected on flat surfaces, not curved ones. With the mild curve ratios of most monitors you probably wouldnt really notice, but once you really get enough curvature that it makes sense, like when using a beamer on a curved 180 screen, you actually need warping software or it looks completely wrong.

So your tv does 200Hz or sorta does it. But you only get 69FPS which makes that rather pointless and which is exactly what I wanted to say about 144Hz gaming monitors. Might be nice for FPS shooters, but anyone who flies DCS and can maintain such high framerates, is not pushing his settings high enough, doesnt have enough monitors or should really look in to buying a VR set ;)

You should try running fraps to show your framerate instead of the built in frame rate counter. I have seen sims where if you enable vsync in game, it clashes with nVidia's vsync and you essentially half your framerate. The game would show 50 or 60 FPS, but in reality you where only getting 25 or 30 FPS. The solution there is first double checking in nvidia control panel that you are not forcing vsync. If that doesnt help either disable it in game and force it in nvidia, or the other way around. Just make sure you dont have both the game and the driver trying to sync to your monitor. Another potential explanation that I have also encountered while playing a flight sim on a 4K TV; in full screen mode, the game let me select a resolution, but didnt let me choose a refresh rate, and selected the lowest refresh rate my TV supported. On my TV that happened to be 24 or 25 Hz. So I got 25FPS. The solution there was not running in full screen mode (but what is called full screen emulation, it still fills the screen but uses windows desktop settings for resolution and refresh rate, same as disabling fullscreen in DCS). Since DCS doesnt have a refresh rate setting either, it may have the exact same problem.

I think you may be confused yourself. Hyperthreading is a hardware feature, that allows a single physical CPU core to work on 2 software threads. It doesnt really execute both threads in parallel, but it can switch instantly between threads without flushing the pipeline whenever one thread is stalled. This is usually beneficial and increases throughput when there are many software threads to work on. But if you have more physical CPU cores than software threads, its rather pointless. And it can hurt performance if the OS is scheduling those few threads on the same physical core as earlier windows versions did, because they didnt distinguish between logical and physical cores. Current windows versions are fully aware of the difference between logical and physical cores and will already do what the OP in this thread wanted to force: schedule different CPU heavy threads on different physical cores. This forum thread (no pun intended) should be locked and forgotten. Its no longer relevant.

Did you disable hyperthreading (in the bios) or did you set an affinity mask? Those are two very different things. Disabling hyperthreading really shouldnt cause any performance degradation on DCS. However, if you set an affinity mask to pin DCS to certain cores, its very easy to mess up and select HT cores that share the same physical core, basically giving you single core performance. My guess is that is what happened. BTW, the original post is from 2013 and for an old AMD processor that used a very different form of hyperthreading. It is long obsolete. Unlike XP, or even 7, Windows 10 does a decent job managing hyperthreading cores. But if you wanted to experiment anyway, you dont need to mess with LUA, you can set core affinity directly in task manager with a few clicks.

When i said comparing a 1070 vs a 1070 Ti would be interesting, I meant comparing their DCS performance, as it may help prove or disprove my fillrate theory. I know how they compare in AAA games.

Where have you seen ANY recent DCS benchmarks comparing various videocards in otherwise identical systems? Or do you mean you have seen benchmarks of other games?

True for the 1080 and 2080. But for instance not for the 1070 vs 1070 Ti. Would be interesting to see how much difference there is between a 1070 and a 1070 Ti

See also this post: https://forums.eagle.ru/showpost.php?p=3897350&postcount=12 It may explain why many people think more vram increases performance. In reality cards that have more vram, like the Radeon 7 or the Ti versions of the 1080 and 2080, often have significantly higher texel fill rates, as they have more TMU and ROPs, which is much more likely the cause of their increased performance in DCS. Games that rely more on shaders than textures, which is probably most current popular games, will not benefit nearly as much from the extra texels. But my guess is DCS does, because its fill rate limited. So yes, you want the Ti version. But not because it has more vram.

A while ago I built a simpit for my gliding club. Having considered everything including multiple projectors, finally I went with a 50" curved 4K TV and headtracker to keep price down. I will say this: the curvature is pointless. It really doesnt look any better than the flat TV I have at home, and being completely honest, the projection/distortion at high field of view settings, is actually worse. Why? Because all games are rendered on a to a virtual flat pane, not a curved one. To get distortion free curved projection, you would actually need some expensive third party warping software. So my advice on curved: dont bother. As for triple screens; i have triple 28" on my desk. And I do not recommend that either. System requirements for running 3 screens, even just at HD resolution are insane. More over, triple wide screen introduces other problems, again with projection. The only way for that to make sense is to have your virtual FoV being the same as the physical one. Thats not undoable, but it does require you either put your screens REALLY close to you and the side monitors, literally, to your left and right. Or you need 3 gigantic screens so you can comfortably sit in the middle. Anything else, again looks weird and distorted. Finally, when flying, horizontal field of view is not all that important. Vertical is at least as important, you want to look up in a turn. Triple monitor doesnt help here. It may even make it worse, as you are inclined to fly zoomed in more. So my advice is again: dont bother with triple screen either. Either go for VR (pimax if you can manage and afford the GPU for), or use a single large monitor (larger IS better) and a head tracker.

Thanks for sharing. We need more info like this. Although I do not think I share your analysis. As others have said, its not likely the VRAM that makes the difference. I have talked extensively with the developer of a (very) different flight sim, but what it has in common with DCS is that its pretty light on using shaders, but uses extremely large sceneries that are basically just big high res textures on a relatively simple mesh and transparent bitmaps for clouds. He says texture (aka texel) fill rate is the biggest GPU bottleneck, at least for his sim. So I looked up some numbers for texel fill rate, and what do you know: 1080: 277.3 GTexel/s 1080 Ti 354.4 GTexel/s 2080: 314.6 GTexel/s 2080 Ti 420.2 GTexel/s People attribute the performance boost of the Ti versions to the extra vram, I think its more likely the extra ROPs and TMUs and resulting massively increased texel fill rate. The above also shows why a 2080 may not be much of an upgrade over a 1080 Ti, it may even be a downgrade. For the record: Vega64 393.2 GTexel/s Radeon VII 420.0 GTexel/s People are excited for the Radeon VII, and that may be justified. But no so much because it has 16GB of VRAM, but because its a texel monster, on par with the 2080Ti. Vega64 may be the most sensible choice though, if money matters and if my presumption is correct that DCS performance is also heavily fill rate dependent.

Im skeptical we can draw any conclusions from what windows reports as vram used. I suspect its a lot like windows task manager showing "memory used" in older versions of windows; it would basically always show ~80% used no matter what. People would think they needed more ram, because they where running "out of memory", but adding more ram would not help, because windows would just use that extra ram for caching more data from disk (since W10, task manager no longer shows disk cache as "being used" so its a bit more meaningful now). With GPUs, the vram usage is not governed by the OS but the drivers. Drivers will allocate vram if they think they may need it, but they may not actually use it all. Windows wouldnt be able to tell the difference. And its also reasonable to assume drivers will load/keep as many textures and other things in vram as they can, even if that data may never be needed (again) or only very rarely. So "running out of vram" might only result in a tiny little unmeasurable performance hit, just like "running out of system memory" (when in reality, it was used as cache) would basically only result in windows reducing the size of the disk cache by a bit, and adding more ram would only increase the cache, which (unlike actually running out of ram and needing to go to virtual memory) might make no difference at all in many scenarios. Adding 16GB of ram was not going to make notepad any faster, even if windows said it was using almost all your memory.

It really cant be caused by disabling hyperthreading, unless something in the background is keeping 4 or 5 other cores fully occupied, but it could be some secondary effect, where disabling HT somehow affects your overclock or turbo boost speed or something else.

Lots of people seek advice on buying computer hardware for DCS, but about the only thing they can hope for is opinions and hearsay. I would like to see some facts. How much difference does ram speed really make? is there a tangible difference between cpu's from AMD or intel at similar frequency? how do otherwise comparable nvidia cards perform vs AMD cards? is a midrange RTX series a better choice than an older highend 10x0 card ? Tiborr did some amazing work back in 2016 using DCS1.5: https://forums.eagle.ru/showthread.php?t=157374 But I cant find anything similar for 2.5 and using current day CPUs and GPUs . DCS is unfortunately rarely if ever used as a benchmark in GPU or CPU reviews, and I dont have 10 different CPUs and GPUs laying around. So I propose doing the following; 1) someone record a track using the Su25 on Caucasus map. Preferably with a fair bit of "action", a mix of low and high altitude, cockpit and exterior views, etc. Everyone would be able to play that track and provide framerate results, they dont need to buy any modules. Hopefully one day some review sites could also start using this? 2) Someone record a second standard benchmark track using a high fidelity plane that is going to be tougher on the system. With mirrors on. Possibly on persian gulf map? I would love to use the F14 for this, but its track recording or playback seems to be broken, so I will leave it others to pick a plane for this. Which one is both hard on the GPU and popular? tracks should be short, below 5 minutes if possible. Once two tracks are selected, we would need to standardize settings, and I think the easiest way is to use the game's own "high" preset (just with vsync disabled) and the most common 1080p resolution. I would use high because its more future proof and probably closer to what we fly or want to fly than the other presets and 1080p because its by far the most common and almost everyone can run at that resolution. Results of this test will not directly tell you if some video card is fast enough to drive the game at 4K 4xMSAA or in VR, but it should give at least an idea if card X is better or worse than card Y And after some time, we may learn that one should achieve at least 150FPS (or whatever) on that bench to have a smooth experience in VR or in 4K. I will gladly compile, analyse and publish results. But I would like someone else to make the tracks, as Im not only a novice DCS pilot, I know nothing about flying the Su25 :).