Gunnars Driver

Very nice to read :-) You can use a graphic card vendors OC-tool for overclocking the GPU. There is other than Asus oc-tool but its the only Ive used( having had Asus cards). They do mild overclocks and has a feature like auto overclock. It tests the card for around 10minutes, the än its done. I think Ive usually get in the ball park of 10% OC. I guess its possible to get higher clocks by increasing cooling and manually clock it, but this is a quick way to a bit more performance.

A parenteris to 8KX: (Strange, I think we have quite comparable hardware. I come from 3 year of Rift CV1 and after upgrading H/W and starting to use the Reverb I think the equivalent to ASW ( motionprojection) is better and as long as the FPS, or more exact; the time to produce a frame isnt longer than the 45FPS time limit( 22ms) there is not even a single stutter or ghosting. Much better than oculus ASW. Still got the CV1 an with the new hardware it also is smooth, mostly but I think this is due to a good overhead in GPU performance.) I was looking into Pimax but was afraid to not getting it work good enough to enjoy gaming in DCS. I know they work far enough in other games I play. The DCS and pimax not running properly got me to buy the Reverbs now, but I feel its only a between Oculus CV1 and next generation of VR with a lot better FOV. Really holding my thumbs for 8KX for the upcoming weeks to see how people manage to make them work with DCS.

It should not be very difficult to make a helo fly according to the charts in DCS. But it is time consuming to test and alter all parameters to make it fit the performance data. You also need a deeper understanding about helicopter aerodynamics and how Density Altitude affects the rotor system( lift drag etc) an how it affects a turbine engine. In most cases the limiting factor is not the same in different altitudes and temperatures. Havent really flown the Huey that much lately, but earlier the EGT was limiting at low altitude and cold weather. That isnt the real behaviour. Also, the NF/NR drooped way to early, not possible to reach the normal torque on the torquemeter. Still, the performance was not that bad despite the engine performing very poor on the instruments. There was another thead about the Huey not flying straight. After a test I found it to fly properly but there is something wrong with the ball, it does not center when flying straight and level, it also behaves strange during banking. I think the coding parameters do not take the right data into count.

Forgot to tell that I actually was able to run DCS with my i7 at 4.9ghz, and with that speed there would be a very small difference to the i9 9900KS@5ghz in DCS. Because of this, if DCS was the only thing, a change of CPU could'nt be motivated. My computer is also used by my wife and being away I didnt dare to leave home with the 4.9ghz setting(fear of angry wife due to not working PC), thus making me set it at a guaranted safe overclock. I guess the i7 4690 should be good for at least 4.7 or 4.8ghz.

Best option for your 4790 would be a good overclock. I've recently upgraded from i7 3770K overklocked to 4.7Ghz, 16gb ram and GTX1080 to 9900KS, 32Gb 3200mhz RAM, GTX2080ti 11Gb. With the same settings in DCS I got somewhere around 15%(guesstimated) more of the time with 90FPS instead of 45FPS. The old hardware almost never got under 45FPS, and with the new and same settings, I newer saw it stutter. Event if the new CPU is some kind of rocket, it only differs 10% in Clock to 5.1/5.2Ghz. I guess the 9900KS is slightly faster per clock but in real DCS gaming, I still could see the single CPU core getting very high load. Maybe average 10-15% lower load, and with the same setting barely keeping below 100% at the worst case scenarios. I dont think you should count on very big gains for the CPU only. I saw better performance but I wasnt the CPU only that gave me these. I do otker stuff than DCS, other games and also workstation so the 9900KS was clearly motivated from that( and also, I hade the Money to spare anyway). If you get yourself a descent overclock, maybe better cooling to cope with the heat and put the rest of the money you can spare on a new videocard, I think you'll get the best bang for the buck. RAM-speed, do not show big differences in gaming. There is a difference but not very big numbers. For RAM yore sitting on, if not very slow it may expensive per gain in FPS. There is a good thread on this forum with tests of different CPU Clock, RAM Clock etc. Id say, when anyhow buing RAM you can put some extra cash in for speed but upgrading same type of RAM to faster is expensive. I also think there is a lot of people around having RAM modules with XMP-profiles but not running the XMP-profile activated in the mainboard BIOS. Do you have that knowledge about your current RAM and settings ?

In your case VR should be better than without. VR lenses is most probable set to let the eye focus on infinity. For me, having past 50yrs, I starting to need glasses to read stuff in cockpit(presbyopia) when flying IRL and most things closer than 1m/3ft is not perfectly sharp....but in VR they is :-) @TWC_SLAG: Did you consult an optician to find out if ou can get any help with the eye vs VR ?

+1. FpsVR is extremely helpfull to find out what is the limiting factor at each time, and also to see how changes in settings is affecting each part. With Oculus before I had a lot of other software running, not seen inside VR, so basically I flew to test, and then quickly took of the VR-googles and looked att GPU-Z, task manager, recourses monitor etc, and tried to imagine lets say 5 seconds back in time, when I hade seen a lower FPS trying to fins out if it was a CPU core max out or RAM max out or GPU max out. In fpsVR you can see it in the googles real time, as a spike in CPU or GPU or RAM happens etc, and you can se the actual frame time for both CPU and GPU. I just updated to new hardware and from Rift CV1 to Reverb. With the help of thuds guide at vr4dcs.com I got started quickly and with very good help from fpsVR the setup for reaching close to what would be possible with my hardware took me a few hours. Without fpsVR I guess it would have taken ten times longer. I also now know a lot more now about how different settings affect performance. Actually, the VR settings for my Oculus Rift was never really finished, despite having had and used them for more than 3 years. FpsVR is king.

I havent tested with both 16bg and 32 with the same hardware but after a bigger upgrade with 32Gb I found out that my rig uses around 15-26Gb of RAM when playing DCS. I guess you can get limited in your settings to stay below 16Gb if you only have 16. A fpsVR frameshot with 24GB RAM usage: https://forums.eagle.ru/showpost.php?p=4170525&postcount=24

For DCS you’re not much helped with a i9 instead of a i5 at the same clock. I still think you should check the load on that single core when DCS-gaming. It might be a eye opener. Before my upgrade and with i7@4.7 +1080 I was limited by the CPU. Seing people writing that Pixeldensity only put load on the GPU, I know that for me I got a lot higher CPU load when using higher PD About the overclock, I cant see that a proper overclock can not change things to the better. The decision not to use overclocking is of course up to you.

Yes, You need the steam even with stand alone DCS(I have it also). And you need it exactly as per the thud guide for getting a good setup in the beginning. Follow the guide just as it says.

Related but Worth an own post, I think: After tweaking my system(I had a lot of help from fpsVR, but I dont Think it works with oculus) I have the following usage: -CPU 70-90% on the most loaded core( less than 10% total) -GPU 70- Close to 100% -Somewhere around 9.5 to 10.7Gb of the 11 available on the 2080ti -Around 16 to 26 Gb of RAM Ive tried to get the max out of the system, as much quality as possible without seing it go below 45FPS. I didnt Think that I needed more than 16gb, nut I do. I didnt Think the GPU would go that Close to the 11Gb RAM available.

Which CPU at which clock? 20-25 FPS is very low. I have used Rift CV1 for 3 years with GTX1080, and never really seen anything below 45fps. The CPU Clock is really important! Cant stress that enough... The overall CPU value says more or less nothing. DCS can only use about two cores and only one for the graphics. My i9 9900KS shows 8% load, but if I look at each core I find the one and only core that actually do the job is somwhere around 70-100%. Look at this: https://forums.eagle.ru/showpost.php?p=4163309&postcount=288 Id say you have to check that core and you have to find out what hw is bottlenecking in different situations. I recently upgraded from a i7@4.7ghz to a i9@5 to 5.2ghz and GTX1080 to RYX2080ti, 16Gb RAM to 32, game directory from classic SSD to a fast NMVe-stick. With the same old settings and the old Rift I maybe got 15-20% more FPS( that’ll say, 15-20% more of the time reaching 90fps.) I didnt use the rift much with the new hardware, did get a HP reverb that has much higher resolution than the Rift S. After doing a proper setup, I newer see less than 45 with the ASW-equivalent setting. At altitude, over sea or looking up always 90fps. With my old setup the CPU was the limiting factor, @4.7ghz. I guess in most systems, if below that 4.7 the CPU will be limiting. In my current system, the GPU is the main limit, but not in all situations.

You really shouldnt compare it like that. VR sets hard demands on your harware. DCS in VR, that means a exponent of ^2 for hardware. For DCS you need a CPU at high clocks. You should worke the way ahead to the best settings for your hardware and also your own preferences. When testing settings, you are get much help knowing what part of the H/W that is the limit. For Rift CW1 I had the CPU as the main limiting factor, making me overclock it to a higher freq than initially. I dont know for the Rift S and I dont know what CPU you run and @ wich clock. It might be the CPU that is holding you back and then a better GPU wont be able to perform as per spec. With the Rift I used the task manager and resource monitor for CPU, and RAM, and GPU-Z for the GPU. The PixelDensity /Supersampling was more needed with older gens of VR like my old Rift CV1 to sharpen the pixture to compensate for the low resolution. I just got the HP Reverb and with these I dont notive any difference in picture quality between PD1.0 and 1.2 or 1.4. It sure loads the hardware a lot though. I run PD 1.0 with Reverb. Rift S may be in between Old Rift and HP Reverb, I dont know but there should be info on the net. Try vr4dcs.com https://vr4dcs.com/2019/09/15/rift-s-setup-in-dcs/and the Rifts S giude. PD will bog down your FPS so keep it as low as acceptable by your means.

Just got HP Reverb, and doing settings for finding sweet spot. Found the print screens I did had 1430x1399 resolution. Dont know how DVS handles the print screen, I had expected them to come out at the same resolution as the set resolution on the Reverbs(2205x2161 or something like that). I have followed Thud's setup on vr4dcs.com. Just wondering if something isnt ritght with the settings on the reverb. What resolution do you get if doing aprint screen with 100% resolution on reverb ? [Edit]Found out. The print sceen is taken from the DCS-window on the regulat monitor.

This is my rig during dcs with oculus rift: ”CPU13” and ”CPU15” is doing all the DCS-calculations. In my case its first thread on core 7 and 8. Most of the other cores(physical or logical), isnt doing that much work at all. Except for windows processes, oculus software, Asus GPU overclocking, a corsair fan software, also task manager, system resources to get the CPU workload and a picture sofrwarw to make that picture was running. When doing normal gaming the other cores would be slightly less loaded. It is very easy to see that when core 15 maxes out to 100%, then the FPS would go below 90, kicking in ASW. It is that core(thread) only.

Yes, and its really easy to check that DCS is very high on one core only and the other mostly no usage. My 9900KS shows a total of 8% load, but logical core 15 is 70-100%. Earlier with an i7@4.7ghz GTX1080 the CPU was the limiting factor for my settings and increasing the CPU frequency made a big inprovement in FPS. Despite this, most of the cores showed low usage and only one core 70-100%.

Just tried some flights. With a gross weight of 7500 Pounds and 8500 Pounds and flying something close to normal cruise IRL, 90-105 knots, pedal released/centered it did fly straight. It showed the ball around half ball(a Little less than half actually) to the left. I tested this by landing and lining up on a raunway, checked the direction gyro and then flew along the same runway, and both use the Visual cue inside and external view. So the Huey itself flies straight and nice with pedals around centered at normal cruise, at least for me. The ball doesnt center, its wrong. For the tests above with the ball centered it flew aorund 5 degree of, i.e nose 5 degrees to the left. At higher speeds/Power the ball goes away more, despite having the Huey flying straight and level. The ball doesnt work like it should in other means also so it should get a fix.

Havent played with the DCS Huey for a while. Don't know if they made any bigger changes lately but earlier it felt quite ok for a game. That said, its quite a lot that isn’t like a real Huey/helo. For real helos: The pedal position vs the airspeed is dependent on the gross weight. There is a big difference if only one pilot, low on fuel and no load vs at maximum take off weight. At higher weight you need more torque for tha same speed=more left pedal. The outer air temperature also make a difference, higher temps make the need for more rotor blade angle -> more rotor drag -> more torque= more left pedal. The tail fin is inclined to make lift(sideways) to release the tail rotor load at higher speeds. Normally, in most helos the pedals are somewhere around the middle at normal cruise speeds.

I did never even try the dcs VR preset, only self tuned. Well, I actually got the Reverb yesterday so last night was all about performance tuning :) Not 100% done but at least 50% I guess.

Ï think its a matter of your own choise. I havent played that much with the UH-1 but I use the UH-1 syntetic trim( I have TM Warthog) as I'm use to trim the cyclic to be able to for example look inside without loosing the attitude. That syntetic trim is a bit dangerous making you loose throw, so you need to take that in acount when using it. IRL Bell 206 Jetranger cyclic is without springs and no artificial feeling, you can adjust friction for the stick. If you make the stick non centering it will work like this(except for the adjustable spring force I guess). Best choise would be a mechanically trimmable cyclic for helo flying, but I guess they dont come that cheap. A friend is building a rig that will incorporate both trimmable and force trim releasable cyclic, thats the reason for measuring stick movement and force gradients.

Yes, I used that tool too( followed the very good guide from Thud at vr4dcs.com). For mee, I have enough spare CPU power to use fpsVR. I just came from 3 years of Oculus rift with a 4.7ghz i7 + GTX 1080. These tools was’nt present but I did use the task manager/resources + GPU-Z to find the load on CPU-GPU during tuning. With the oculus rift it always looked like the CPU was bottlenecking, even after the upgrade to i9@5.1ghz an RTX2080ti. The single CPU core running DCS goes quite high. With the Reverb the GPU is higher loaded and the CPU is “in green” in fpsVR at all times.

Got my Reverb today and after setup i have steady 45fps at low level fast speed. 90fps is more rare, looking up of flying high. 45fps came with a setting like the oculus ASW( dont remember the name on it). Beforethis I had 75-90 fps at all times but not a perfect flow, which that setting got me. Found out via a software( fpsVR) that with Reverb its the RTX2080ti that is bottlenecking on my rig, so I made a higher overclock on the GPU.

Full Control throw in a real helo is around 20cm both in pitch and roll, from stop to stop, measured at middle position on the cyclic handle.( = from center position to stop, around 10cm) I measured the NH90 the other day for a friend building a DCS-rig. A long extension might make control range more than IRL.

Here is my settings. We have similar H/W. I get mostly 90FPS and never below 45. (I have a i9 9900KS@5.1Ghz on all cores, 32Gb@3200Mhz RAM, RTX 2080ti 11Gb.) First step: - Have you setup tour BIOS to get your i9 to always stay at 5ghz during gaming? The Intel standard for ’turbo’ is time limited, but with the right settings(and a good CPU cooler) you kan have it do 5ghz all the time. Also, if it doesnt do 5Ghz on all cores, either make it does or make sure DCS uses the core that actually do 5ghz. ( there is info on how to perform that setting in this forum) As said earlier in this thread, DCS do mainly only use one core. Make sure that that core is speeding. [Edit]My VR settings. I've just upgraded the rig so settings is copied from the old rig. It might be possible to use a little higher settings but it is mainly the CPU single core performance that is limiting and I have increased that part only(I Think) by roughly 8% by going from 4.7 to 5.1Ghz so there isnt room for very much higher settings. I have used Pixel Density 1.6 to easeier spot other Aircrafts, in the screen shot 1.2 was set for some tests. PD1.2 gives med more or less 90FPS all the time, never ever below 45. PD=1.6 gives 90FPS most of the time, and 45 at very low level/high speed over Heavy Graphics like a lot of Threes and stuff. Not seen it go below 45FPS. The 1.2 setting give me a fantastic smooth ride all the time and the 1.6 have some minor degraded "flow" sometime. The preset buttons down right shouldnt be used when copying these settings. Yes, I have a ASUS RTX 2080ti OC card, I am using asus stock overclocking tool and by one button click to slighty overclock the RTX. I dont know if it actually makes any difference on DCS in VR(Think it doesnt), because the CPU is the limiting factor. In other games it does, at least with the old GTX1080.

DCS have very high demands on the CPU single core performance for VR. When a 2080ti gives good gains over a 1080ti in other games, you could be sitting with a CPU that limit the DCS performance by single core calculations. In this case you might get a smaller difference in DCS when upgrading the video card. I just when from i7 3770K@4.7Ghz/16mb/GTX1080 8Gb to i9 9900KS@5Ghz/32mb/RTX2080ti 11Gb and with the same settings as before its a bit more prone to keeping 90hz over a flight but the difference isnt huge. I think this small difference comes mainly from the increased clock giving the CPU a greater single core performance. The GTX1080 didnt up at 100% load. I think its good to be aware if the DCS CPU demands setting the limit for performance in VR. ( never really used DCS without VR so I dont know that part).