

rubbra
Members-
Posts
82 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Everything posted by rubbra
-
[2.9.8.1214] Apache and Phantom unflyable in VR
rubbra replied to rubbra's topic in Game Performance Bugs
Yes, it's definitely VRAM - textures to medium makes it go away. But (not a rant at you, just observations) - a) why has that changed so much between patch versions, and why only for AH64 and F4? b) why is it only with SteamVR with PicoConnect? I would understand it if it was still a problem with SteamVR as OXR runtime + VD, but it works just fine. VDXR vs SteamVR with VirtualDesktop is near enough identical in performance. c) The point of using PicoConnect is because it has fewer compression artifacts/quite noticeably better image quality compared to VD d) I know a 4070ti is *only* a 12Gb card, but it's still in the upper end of cards for VR. Do I really have to run it at low terrain and medium cockpit textures just to make it playable with a single unit? Seems crazy to me, especially after all the "optimized" vram usage items in recent patches. -
[2.9.8.1214] Apache and Phantom unflyable in VR
rubbra replied to rubbra's topic in Game Performance Bugs
Further testing - I wondered if it was any OXR layers; I disabled everything, including quad views. No improvement. I don't understand what is specific to the AH64 and F4 that isn't specific to the other modules. The only thing I can think of is VRAM usage; on the occasion that I can manage to manipulate a mouse cursor enough to quit the mission rather than force quit DCS, the fps recovers when back in the "debrief" screen, but only after a couple of seconds. -
[2.9.8.1214] Apache and Phantom unflyable in VR
rubbra replied to rubbra's topic in Game Performance Bugs
Edit - this appears to only affect PicoConnect + SteamVR (as the OXR runtime). VirtualDesktop, VDXR, Wireless - 72fps VirtualDesktop, SteamVR, as XR runtime, Wireless - 72fps PicoConnect, SteamVR, USB wired, 5fps PicoConnect, SteamVR, Wireless, 5fps In both cases, resolution is 3188x3188 or thereabouts (3200x3200 in VD's case, but what's 144 pixels between friends?) -
Both the F4 and Apache have become unusable for me in the most recent (Oct 2024) patch. Point of reference - Caucasus "Takeoff" instant action for F4 gives me 2 fps. Caucasus Cold and Dark IA for Apache gives 3-5 fps, vs flying pretty much any scenario, multiplayer or single player in the FA/18 or MI-8 or MB339, where I get 72fps. As a test, I created a new mission on Afghanistan. Placed a single UH-1 at Kandahar. 72fps. Change that unit to an Apache, 5fps. OpenXR, SteamVR as OXR runtime using PicoConnect, USB wired to a Pico 4 over USB3. Running at 3188x3188 MSAA 2x 4070ti, Ryzen 9 3900x, 96Gb RAM Additional/non-standard stuff - TacView and SimHaptic - but neither of those are new.
-
Go to c:\users\you\saved games and rename the DCS or DCS.openbeta folder to DCS.old or DCS.openbeta.old. That way you can revert back by renaming it back.
-
DCS World 2.9.5.55918 Constant Frame Time Spikes
rubbra replied to FierceLV's topic in Game Performance Bugs
I've removed lua hooks for both "Bergison's Moving Map" and "DCS Web Editor API", and this spike has gone away. Might be worth exploring for others. -
I have also got this issue, but found that a clean saved games folder cured it. Now trying to copy things back item by item and testing after each change to track down the cause.
-
DCS World 2.9.5.55918 Constant Frame Time Spikes
rubbra replied to FierceLV's topic in Game Performance Bugs
I see a 120ms frame roughly every 2 seconds, in VR, but only in multiplayer on Caucasus. My test case is to slot into a quiet MP server (1/2 players) at Kobuleti in the A10C-2 (though it does it in other modules). This reliably produces the periodic spike. Slot into a Syria server, same conditions and module, no periodic spike. I've tried lowering the mouse poll rate as suggested above, no difference. -
I know this is resurrecting an old thread, but has anyone encountered SimFFB (incl the version linked a couple of posts above) randomly causing *extreme* violent shaking of the FFB device (Sidewinder 2 + extra resistors in my case) ? It'll typically do it on startup for me.
-
Why is there a plan to implement right hand traffic in Cyprus? Has it changed really recently? I'm pretty sure it's left hand traffic... (e.g. https://www.cyprusparadise.com/articles/practical-information/ and https://www.enjoytravel.com/en/travel-news/driving/driving-in-cyprus-15-things-you-need-to-know and https://www.worldstandards.eu/cars/list-of-left-driving-countries/ ) edit - I see the description has now changed to say left hand traffic.
-
Parking brake on. That’s all you need. (And no rearming or control input to move the aircraft)
-
Last night I was flying WSO with a friend. Ticking along at a solid 40fps in VR (locked/ASW). He turned on the A-G radar, frames dropped a bit. Turned on HRM, I get black screen/hour glass for a good couple of minutes, Oculus complaining that a hardware error had occurred, couldn’t track my movements etc, until he changed the MPD away from the radar; even switching modes to GMT didn’t help. As soon as the radar was off his MPD, back at 40fps. Now I’m not on an amazing CPU - i5-8400 - but I am amazed that the front seat turning on the radar would make my session stop. I get a bit of stutter in single player when I turn the radar on, but MP appears to be a “just say no” for the A-G radar for me. Any tweaking tips?
-
Annoyingly I didn't get a screenshot, but I just ejected (as WSO) over Syria, then started to have a walk around on the ground. My WSO's shadow had no head, from any angle.
-
I am not 100% sure, but based on a limited amount of testing, I believe that clients joining a server with F15Es on it require more RAM than those without F15s. e.g. tonight I joined Grayflag Syria with 31 players on, Windows killed DCS because it was using too much RAM. I profiled it, and got to 36Gb of memory allocated just at the Slot screen for Grayflag, after a reboot. I then joined another server with no F15 slots/players, 27 people on it, Syria, also after a reboot, 26Gb of memory allocated, a whopping 10Gb less. I've done similar with Sinai maps and found very similar numbers. The obvious test would be to make a server running a mission with an F15 slot and something else, e.g F16 and have two people join, measure RAM usage. Then repeat without the F15. But as I don't own the F15 and don't fly with anyone that does, that's a hard test for me. MT, VR, Rift S, 32Gb RAM, 1080ti. Never had DCS be killed for using too much RAM before. Grayflag Syria on the last release was using about 24-28Gb for me - I happen to know because I was testing the impact of visibility vs textures.
-
I'm on a lowly Rift S, 1080ti, i5 8th gen, and it's the best performing map by far, with the odd stutter here and there low over cities, but I get that on every map. The growing tree thing mentioned above is to do with visibility distance; the higher you have that, the further out the trees grow. Applies to all maps. Edit: wanted to make it clear that Caucasus suffers from the same inflatable trees issues, it’s not just ugra maps.
-
Can reproduce at Ramon too (F16 Instant Action Cold Start), but not all trees; palm trees seem immune.
-
Behaviour - in VR, tops of trees rotate with your head. Reproduce: In VR, Go to Instant Action, F18, Free Flight. Look to your left, move your head left and right. The tops of the trees rotate. I have tried to capture a video, but this behaviour doesn't show in the 2d desktop VR mirror. MT version, OpenXR, Rift S.
-
In Vr by any chance? I have also noticed a huge performance difference of BS3 vs pretty much anything else. I’ve been flying quite a lot of BS2 in 2.8 recently, constant 40fps (ASW locked), in all scenarios, including the last thing I did before the update came out, which was the Caucasus instant action shooting range. Do the same with 2.8.1 and BS3, and I’m struggling around 14/15 fps. I had assumed it was just BS3, so gave up on it in disgust assuming it will improve in later releases. I hadn’t considered it might be BS2 as well - will test tonight.
-
I’ve recently found that Openxr toolkit (or perhaps opencomposite - the combo certainly) causes that behaviour on all modules and also constant micro stuttering in MSFS and il2. I’ve disabled it for now
-
Not entirely related to boresighting, but still IHADSS + VR... Do any other VR users experience the IHADSS not quite being central (which I guess might be true to life?), and thus having to choose whether to have ALT or IAS in focus? On my Rift S, with default VR "IPD" in DCS, the Altitude is always slightly out of the "properly rendered" circle of sharpness. I can fix it by setting DCS' IPD to 75 or higher, but that's not ideal, or by shifting my HMD off centre, which is uncomfortable.
-
(I hope this hasn’t appeared already - couldn’t see it!) Problem: George’s target list is hard to see Oculus Rift S, using Oculus runtime (i.e. not SteamVR), just in case that makes a difference. Monocle set to right eye/default. George interface appears fairly high in my FoV - would prefer it a bit lower, I think, but not a big deal - which is fine, nice and visible. Tell him to slave to what I’m looking at, and then he produces a target list. Which is nigh on impossible to spot/see as it overlays (mostly) the bottom left of my view, often over the dark cockpit instruments. Sometimes it appears on the right. Not sure why. Also, I see that list double, unless I shut an eye. Workable, but probably not the optimal solution for future releases.