Jump to content

Quip

Members
  • Posts

    245
  • Joined

  • Last visited

Posts posted by Quip

  1. Quest 2-players...
     
    Does anyone use Airlink or are you all using Link cables? And if you're using Airlink, do you too experience problems with the Airlink dropping from time to time?
     
    When that happens, DCS isn't able to recapture the graphics and I end up having to quit the session. The Headset will start tracking etc again, and I can see the head movements on the PC monitor, but inside the VR the image is black.
     
    Anyone experiencing this and is there any way to get the game to "recapture" the graphics or to prevent this from happening at all?

    (Other games, such as MSFS, can recapture the gfx easily when this happens)
  2. I have a new problem plying on Quest 2 and 3080 gfx, win10, lots of RAM etc. I have only seen it in multiplayer games, but I have very little playtime in the latest openbeta so I can't be certain that it's only in MP or not.

    I get stable 36fps _entering_ the mission.
    I press F10 and then F1
    I get 6-12 fps.

    So far I haven't been able to get out of this situation.

     

    So something's happening...

    I'm going to wager a bet. Lately in DCS, the F10 map is slow on initial load. It takes several seconds for it to load the textures (even in "2D mode"). But this is only true the first time you load it. Thereafter it stays cached (for the most part). I bet what's happening is that when you go to F10, the map debuffers the world and takes its place so to speak. So when you get back to F1, the game plays catch up, trying to reload the world but failing to. Partly because it just can't do it and partly because the F10 map is taking up its place in memory.

    This is a guess, but an educated guess and I wouldn't be surprised if at least part of this is true.

  3. On 4/20/2022 at 1:48 AM, chfeng said:

    I do not think so(not gpu or headset issue) 'cos, I am playing lots of game with my rig and set both in vr and non-vr without any problems. Even with the latest games. I believe, somewhere deep in DCS VR, some code fucks up very bad. 

    Even with the highest system demanding games, I can play for 10 hrs non-stop w/o any problems.(while gpu is %99~100 at all times)

     

    I don't know if you said this as an answer to my reply, but if you read my post you will see I say the same thing: it's only in DCS, and I can hapily run any other game at full draw on the gfx card without any similar problems.

  4. 17 hours ago, Delta59R said:

    Does it occur on the home screen or only while in play? Are you just experiencing low frame rate or are the frames being dropped as well?

    I can push it to occur in the menu - but that takes really over-driving the gfx settings in the occulus and such. So I would answer "it's only in-game for me", as there is a limit to what you should expect to be possible in the first place.

  5. Yup! Precisely!

    This is what I see too (although I didn't mention the re-projection errors as those are an artifact of the other problems.

    See here!! (Preferably, delete this thread and add it to mine? Also Uprate my post to make it visible)

    Edit: It's a good thing you're on a different VR set (I have Quest 2) and a slightly different Gfx card (even better than mine, I only have an RTX 3080). This proves it's not a VR kit or specific gfx card issue.


    PS: I'm sorry you have it too, but glad I'm not alone...

     

  6. Easy. That took all but 5 minutes.

    The Afterburner view is in an image.

    And attached a .7z with the

    • full log
    • the DCS .crash file, as well as the .dmp file
    • two of the SRS logs

    Since the log uses UTC rather than Zulu, the time to look for is 18:52:30. I've pasted the portion from 18:51:25 till I quit the mission below.

    2022-04-12 18:51:25.949 INFO    NET: Current ping: 37.6ms
    2022-04-12 18:51:55.961 INFO    NET: Current ping: 36.8ms
    2022-04-12 18:52:05.932 WARNING Mission: register_unit(1005810,18188033) : db.units['1005810'] table is missing
    2022-04-12 18:52:25.967 INFO    NET: Current ping: 36.9ms
    2022-04-12 18:52:56.029 INFO    NET: Current ping: 36.7ms
    2022-04-12 18:53:18.227 ERROR   SOUND: invalid source_params(woLA-16800524:aircrafts/f-16/f110ge129_rpm3_front): gain
    2022-04-12 18:53:18.227 ERROR   SOUND: invalid source_params(woLA-16800524:aircrafts/f-16/f110ge129_rpm3_side): gain
    2022-04-12 18:53:26.039 INFO    NET: Current ping: 37.0ms
    2022-04-12 18:54:23.246 WARNING LOG: 1 duplicate message(s) skipped.
    2022-04-12 18:54:23.246 INFO    NET: disconnected: 0 
    2022-04-12 18:54:23.256 INFO    NET: disconnected from server
    2022-04-12 18:54:23.732 INFO    NET: client has stopped
    2022-04-12 18:54:23.864 INFO    Dispatcher: Stop

     

    Logs.7z

    RTX 3080 with Log.png

  7. Hey

    I'm still chasing a very odd problem that I'm having.

    I got my Zotac RTX 3080 "Gaming" a few days ago, which has replaced my 1080 Ti.
    I have used DDU to remove all old drivers and installed nVidia from scratch.
    The card is mounted with 2x 8-pin cables, as you can see here.

    My problem however, is really really odd.
    While playing DCS in VR (which is what I've used basically all the time since getting the new card as it's here that I have the problem).

    So what is it? Well...
    As you see from this image, the card will suddenly do something which doesn't make any sense at all. It will:

    • spike the GPU at 100%...
    • while lowering (!) power consumption by 1/3
    • while increasing (!) demand on the Bus
    • and the result is a net loss of fps by 2/3 (from 45 to 15)

    (Disregard when the graph shows 0 fps as that's when I have the game in the background, those moments are "normal".)

    It will stay like the indefinitely, or until I pause the game--which doesn't pause the 3D rendering and VR! After a 5s or so in pause and task switched to a different program, the curves will bounce back and I can go back to the game, for a while until it starts again.

    I have also removed all Oculus software following the correct procedures, and reinstalled that too.
    I normally use the Quest 2 via a USB cable which is more stable than going by Wi-fi. But I have also tried over Wi-fi, and I have exact same behavior there too, i.e with the VR headset having zero cables going into the PC!

    I've stressed the card running MSI Kombustor for 15 min, and it's stressing the card way beyond anything DCS does even in VR, with higher average GPU % load and much higher average Power (W) usage.
    I've also started OBS and some other tools simultaneously to push the card even further.
    And it's doing this without a hiccup.
    Here's an image of this test

    I have also run DCS on "pancake" (i.e, not VR) and have zero problems when used like that. The forums are void of this problem as far as I've been able to find.

    Even though I really don't think the problem is related to my PSU, just to be sure, I have today replaced my old Corsair CX750 with a brand new Cooler Master V850 Gold V2. And I still have the same behavior...


    I would normally expect the W to go hand in hand with the GPU %, but it doesn't. And the behavior is completely random; I'm not doing anything special, or going to a "dense" place. It can happen with me sitting there not doing anything. That means that the load on the CPU and/or GPU aren't changing.

    I have tried bumping up the fans to max to see if it was correlated with temperatures, but nope.


    Does anyone have a clue as to what this is?

    Anyone? Anything?
    Help is MUCH appreciated.

    • Like 1
    • Thanks 1
  8. ADDENDUM

    The problem seems to be linked to the altitude of the target (area). Above a certain altitude, the AI will not attack, even if the terrain is perfectly flat. Let's agree that a lake is perfectly flat.

    1. In the first image, the lake is at 2077 m, and the AI will not attack
    2. In the second image, the lake is at 2077 m, and the AI will not attack
    3. In the third, the lake is at 1078, and the AI attacks as expected

    I have tried many more altitudes, and 1462 m doesn't work.

    It seems to be tied to altitude somehow. But it's all very random. It may be linked to the altitude at which the AI is looking to drop the Bk90's.

    • As the area nears 1200 or 1300 meters, the AI starts behaving erratically. In the 1148 meter test, over a lake (img 4), the attack is done very late, like under 2 km range, as opposed to the normal 7 or 8 km over flat terrain.
    • Sometimes, the AI won't drop both Bk90's (which it's set to in the mission).

    Heatblur should look into this.

    bk90-2077.jpg

    bk90-1669.jpg

    bk90-1078.jpg

    bk90-1148.jpg

  9. Hey.

     

    I just noticed I'm getting these artifacts which looks like the side panels are pixelated, with the VR Shaders. Yes I have tested with and without and they occur with the shaders only. And the effect is visible in VR too. I only took the screenshot in 2D because it's easier.
    I only see it on some textures and in some aircraft and under some specific lighting conditions only, mainly when it's dark.

    If you you look at the image you can see the effect on the side panels including the switches. But nothing on the center panel.

     

    Anyone else? Any work arounds?

     

    image.png

  10. THANKS!

     

     

    On 10/25/2020 at 10:00 PM, Taz1004 said:

     

    I'll try to explain but if you don't understand, just try the values I suggested.

     

    First, there're two scales. 0-255 and 0-1. Which is what adds to confusion. epoch who asked the question to me was talking in 0-255 scale. But HUDinit.lua uses 0-1 scale. materials.lua also uses 0-255 scale which is another confusion.

     

    Any value between 0-255 (or 0-1) will not work for HUD readability. This is because DCS uses blending mode similar to overlay in Photoshop. Any value above 0 will be added on to background but if the background is already white like against cloud or horizon, it wont get whiter.

     

    If you go negative, the value will be subtracted from white. And make it darker than white. Letting you see against white background.

     

    This wouldn't be the case if they just used normal blending mode and adjusted through opacity. But we're told this is the way it is.

     

×
×
  • Create New...