X-Plane 11.50 Beta 17 is out now – full release notes here. There are a lot of bug fixes in this beta. We are getting near the end of the 11.50 run and the fixes are becoming smaller and more targeted as we try to lock things down.

We fixed several performance issues on the GPU and CPU, on both platforms. Probably the biggest single fix is Mac cloud performance with Metal and no-anti-aliasing; it turns out that having the rendering surface shared between Metal and OpenGL on a Mac puts it into a layout that slows down the GPU. (This is not an issue on Vulkan.) We’ve fixed this by using separate VRAM for cloud rendering vs plugins; GPU performance with Metal should match OpenGL.

VR users: water reflections are fixed, as well as hopefully the crash on quit with Nvidia + Rift headsets, and we’ve tried to fix the VR mouse not being available until a controller is activated on WMR headsets.

Multi-monitor users: we finally figured out why the horizon line wasn’t lined up – it turns out each monitor was using a different monitor’s height and chaos ensued.

If your bug number isn’t listed in the fixes, you don’t need to tell us it’s still broken, but if we did list a bug as fixed and you still see it, please file a bug and include the XPD number if you know it.

Oh, and Mac Catalina users: you no longer have to reboot your machine after an update. (This was an installer issue – the new installer rolled this week.)

Third party developers: Thomson made this survey form! Basically we’re trying to decide where the best place(s) are to discuss future dev, so we figured we’d get some feedback before proceeding. We’ve had some discussions of future extensions to the SDK on Slack, but not all third party devs have time to monitor a Slack channel, and we don’t want to leave people out.

About Ben Supnik

Ben is a software engineer who works on X-Plane; he spends most of his days drinking coffee and swearing at the computer -- sometimes at the same time.

155 comments on “X-Plane 11.50 Beta 17 – Performance Fixes

  1. Is it better/required to have latest nvidea driver 451.67 to run b17? This latest driver seems to be locking my computer with the msi afterburner

    1. I’m on the latest drivers and didn’t have any issues with b16, so I assume b17 shouldn’t be any trouble.

      1. Anything after nvidea 446.14 keep getting locked up with msi afterburner or any other software for oc. I have not overclocked gpu but just trying to adjust fans.
        I7 9700 rtx 2080

        1. I keep on hearing about the issues with the new drivers and MSI Afterburner itself lately, maybe it’s just incompatible, at least in it’s current version?

          1. I would certainly suggest that you start with NOT using MSI – if the sim works without it you can add it and see if it works. These kinds of non-plugin third party ‘things that mess with your GPU’ are something we have _no_ insight into, so we cannot say if they are compatible.

          2. Is there anyother way to increase fan speed for cooling gpu without using msi afterburner or other third party software? Thanks guys..thanks ben for your work

          3. Well, not to be Rude or anything but the MSI plugins should be researched into by Lamar and the X-Plane team. They come standard with any gaming PC from MSI now.

  2. This is the first 11.50 release where Metal is actually faster than OpenGL on my machine.

    1. Sadly not for me. I’m still getting better FPS with OpenGl. And with Metal no clouds with SkyMaxx.

      1. I had no clouds with sky maxx until I updated it, but it could of been down to the fact I’d added an extra ssd and and extra 16gb of ram, as it strangely thought I was using a different pc. I was forced to “freeze” prevoiusly used computers. I’m on Windows but it might be a similar issue?

    2. Not so for me on my iMac 2017. I’m still getting better cloud performance with OpenGL. A quick test sitting on the KSEA ramp with the default Cessna, forward view and using the weather presets:

      OpenGL
      clear sky: 50 fps
      broken: 44

      Metal
      clear sky: 67
      broken: 30

      So Metal outperforms OpenGL as long as there are no clouds. With clouds, the performance is up for debate.

      To be fair, there is a certain randomness to the broken cloud coverage between tests. I would need to run a test using the command line parameters to get a comparable cloud coverage, but I’m definitely not getting better fps with Metal in clouds.

  3. Nice update, probably the overall best Vulkan beta in VR so far for me (Vive Pro and Oculus Rift). Still same Oculus Rift shutdown bug (task manager to kill is still required). No big deal for me since I mainly use my Vive Pro for all my sims and this shuts down fine. Good job Laminar, many thanks!

    1. Hi there, same “task manager” problem on exit for me since several betas (can’t remember exactly which one).
      This happens without VR (never used it) and on Ubuntu 18.04.

      Should I file a bug report?

  4. Thank you all for this update.

    I did notice a couple of this do :
    FPS in VR seems to be stuck to maximum 30fps, no matter what you do (tried it on a couple of different airfield with different aircraft)
    When X-Plane loads and during flying, maxis out 4 cores of the CPU, it doesn’t seems to do anything with other cores (in 11.50b16 this was better balanced)

  5. Still works loverly on Mavericks with each beta getting better and better. At some point I expect it will not.

  6. I see slightly higher FPS compared to b16, also FPS drop when panning view seems to be gone. Nice job dear LR team!

  7. Now that’s a beta release! I’m sat in the updated EGLL developed by Janov, with Orbx True Earth Great Britain in use, in the Zibo 737, with the following settings:

    – Visual Effects: High (HDR)
    – Texture Quality: Maximum
    – Antialiasing: 8x SSAA+FXAA (highest quality)
    – Anisotropic Filtering: 16x
    – Number of World Objects: Maximum

    I’m getting a steady 36 to 37 FPS, panning is smooth and it is stutter free. That is some going. On previous betas I found the need to decrease AA and AF in order to maintain that frame rate, but on beta 17 it’s all maxed.

    I have to add that this is in clear weather and I’m sure with clouds on I’d need to lower the settings, but when comparing FPS between betas I always load the simulator with the same criteria (EGLL by Janov, stand 218, Orbx TE GB in use, Zibo cold and dark, daytime and clear weather) and I’m seeing gains on beta 17 using those settings. I’ll test with weather later, but I’m very happy.

    What is on the roadmap once 11.50 is locked down? I know that Austin mentioned the need for better scenery and weather for a next gen simulator – will we see further improvements to X-Plane 11 or should we be looking forward to X-Plane 12 now?

    Good work guys n gals!

  8. XPD-10898. Vast improvement after deleting preferences.
    However on max objects with a lot of autogen, I still get a dip to sub 20FPS when changing the views but only not longer than a second. Great effort overall.

    1. It would probably be useful to get a new ETW trace – object paging is much better than it used to be but a third party scenery can always have more.

    2. Just to be sure before I do something wrong here, when you say deleting preferences, you are referring to deleting the \xplane\resources\settings.txt file?

      Thanks

          1. I tried to simply rename the “preferences” folder and started X-plane.
            After that, about 15 minutes of black screen (not freezing though).
            Then I closed X-plane, restarted – and my frames went from 35 FPS to 50 FPS.
            Before that, I couldn’t adjust anything to change except “lower frames”.
            I think everyone who hasn’t tried this yet should try this out or simply make a fresh install of X-plane.

            50 FPS with settings:
            Visual Effects: High (HDR)
            Texture Quality: max.
            Antialiasing: 2x SSAA+FXAA
            Antisotopic filtering: 4x
            Base Font size (yeah, I need better glasses 😀 ) : 16px
            Monitor: Full screen simulator
            Lateral Field of View: 108 degrees ( less will give you even more FPS)
            Number of World objects: High.
            Reflection details: low.
            .. enabling Vulkan of cause.
            RTX 2080 , AMD 3900X, SSD, 32GB ram.

          2. Tried it here, and for me was quite the opposite…
            With my current preferences I get 40-45FPS (in VR)
            by deleting the preferences folder, re-configuring all the Graphics to the same setting that I had before it drops to 20-25FPS…
            So, I’ll keep my old preferences, I’m already happy with them.. but thanks for letting me know where to do this.

  9. Thank you guys so much for your effort! I really deeply appreciate it. Just having fun with Xplane in VR AND WILL HAVE IT FOR A LONG PERIOD OF TIME 🙂

  10. This comment is a bit out of scope, but i was tempted to share this demo for unreal engine 5.0:

    https://www.youtube.com/watch?time_continue=523&v=qC5KtatMcUw&feature=emb_logo

    https://www.youtube.com/watch?v=9PmjQvowfAI

    Yes we know x-plane does not (and will not) be using game engines, as it has it own 3d world engine .. however, the reason for bringing this hereof, is for purpose of proving that with modern 3d api’s on new to-be-mainstream hardware and real-time hard-ware rendering, the results are amazing even for open worlds, which tempts one to think how far x-plane 12.0 can go, especially given that it is time for flight sim to zap into next-gen scenery …

    The bottom line, key decisions can be made to put foundations of next-gen simulations both in terms of physics/graphics and AI !!!

    Best regards

  11. Thanks Ben & team. I am on Mac and can confirm the sim is much more usable now. I’m getting 30-40 fps with scattered clouds, high textures and fly Tampas Vegas scenery.

  12. update- even better, I’m getting 38fps with storm clouds at dusk with a down pour with the default 737 & DD Chicago scenery & ORD. amazing update for Mac. now the key is to NOT revert back to issues but just improve it even more!

  13. I have not seen elsewhere about two comms & two navs availability. Will that feature be included? Also, will the Garmin 530 be replaced with a Garmin GTN 650?

  14. VR mouse cursor isn’t working/showing at all. WMR controllers are detecting but there are no bindings so can’t use trigger to click around menus or configure anything, let alone play.

    1. I also have this but only when I start the game through SteamVR environment. When I start the game in the WMR environment, the controllers function correctly.

  15. Lost Plane command in 17, Button push shows active but no sound clip in Troubleshoot. Any ideas?

    1. I’m the author of PlaneCommand.

      PlaneCommand should work just fine in b17; a bunch of other users have been using it with no new issues. Feel free to email me (lee@planecommand.com) with your Log.txt attached and I’ll help you troubleshoot.

      In the meantime, I’d recommend updating to PlaneCommand 3.2.4. I just released it, and it fixes a number of bugs.

  16. Great to see the hard work paying off. Thank you for all your efforts, it’s looking great. Noticed a couple of things in this version. I’ve got a laptop with a rtx card and a intel onboard gpu. It looks like the intel is doing all the work and the rtx is doing nothing. Noticed a couple of times when i change planes mid flight and start a new flight it crashes. i’ll log a bug report for them tomorrow.

    1. I have a 2080ti and it barely even heats up, or maxes out the ram, i think with this sim it is mainly cpu bound. You can get away with using very low end cards and get the same performance.

    2. Drew, go to your Nvidia control panel 3d settings and make sure the top line has selected Prefer high performance nvidia gpu. Also, if you are still using OpenGL instead of Vulkan then look further down and make sure OpenGL goes to your nvidia gpu.

  17. Cheers Ben,

    I don’t normally post on here, but I can see how much effort this has taken, it’s incredible that you’ve all done it as part of X-Plane 11. As a fellow developer in NET and C++, but an armature compared to the X-Plane team, I feel some of your pain!

    I read a quote from a 2018 Microsoft blog recently which I think captures part of your challenge perfectly.

    “With Windows 10 alone we work to deliver quality to over 700 million monthly active Windows 10 devices, over 35 million application titles with greater than 175 million application versions, and 16 million unique hardware/driver combinations.”

    “16 million unique hardware/driver combinations.” What the actual &*%$! And that’s just Windows 10!!!!

    An incredible feat, to get anything working this well, across such a plethora of different devices!!

    P.s. I have also perfected the skill of drinking coffee and swearing at the same time. Also, the number of fixes that come to me in the shower is amazing. Developers should definitely spend several hours a day in the shower 🙂

    1. Mark – you could be a more professional software developer by carrying less current and not rotating so much.

      1. I got it. 😉 Dont put the guy in a spin….perhapscwe should field some if his questions so hes not in a state of flux. 🙂

  18. Great work, it gets better and better! Have you got to the bottom of the “purple smoke on landing” bug yet?

  19. Great work LR team! In B16, with frames locked at 45 fps in NVidia control panel, locked VR ASW to 45 in Oculus tray tool, with super sampling at 1.5, got a solid 45 with only minor stutters. Grapics settings:

    Visual Effects – High HDR
    Tecture Quality – High
    Antialiasing – 4x
    Anisotropic Filter – 8x
    World Objects – Max
    Reflection Detail – Low

    In B17 with same settings, taking a 15 min flight, could not detect any stutters, moving head around in VR, everything looks great!

    I also use PlaneCommand, (awesome plugin for VR), it worked perfectly. I would suggest check your default microphone in windows is set to proper mic.

  20. Im getting a little scared that 11.50 will be complete soon because the last 2 builds (16 &17) I can’t fly as far as I could in March. With all my settings set to low except for the one that does HDR lights it crashes at about 3 hrs into the flight. KLAX-KHOU crashes about 45mi from KHOU. It finaly generated a crash report after a few tries today so it is filed. Its funny that weather my settings are set to high or low, the FPS are about the same.

    1. This is interesting! I did a flight from Las Vegas (KLAS) to TNCM (St Maarten) and near where I crossed from Houston FIR to Jackson Ville FIR I had a crash to desktop.

      An auto crash report was generated and sent so I did not check the log file.

      I resumed the flight using the rsuming feature of the ToLiSS A319 and continued a few miles before the crash position – this time without problems.

      I do not know if that crash to desktop was related to any b17 issues. What I can say is that I did not have any problems with the b16.

      Beside this crash to desktop everything was super smooth, almost no micro stutters (unless panning around fast at the new FlyTampa Las Vegas scenery at 4K resolution, objects and textures maxed out, FXAA and SSAO active, anisotropic filtering 8x, reflections low – but this scenery is super super heavy duty for the PC!).

      Regards,
      Marc
      ——
      CPU: Intel Core i7 9800X (OC 8x 4.8GHz) // CPU Cooling: NZXT KRAKEN X72 // Main Board: ASUS TUF X299 MARK 1 // Graphics Card: GeForce RTX 2080 Ti ROG STRIX O11G // RAM: 32GB (G.Skill Trident Z RGB Series, DDR4-3000, CL 15, quad channel) // Hard Drive: Samsung 970 EVO NVMe SSD 1TB, PCIe 3.0 (M.2) // Optical Drive: LG BH16NS55 // Power Supply: Corsair HX850 High Performance // Head Tracker: Track IR 5 // Mouse: Logitech G203 // Joystick: Thrustmaster T.Flight Hotas Stick X // Keyboard: Logitech G815 (tactile) // Monitor: LG 27UL850-W (3840×2160) + LG FLATRON IPS235 (1920×1080) // OS: Windows 10 Pro 64bit (v. 2004) // SIM: X-Plane 11.50 b17

  21. thanks for the update however Same problem as all prev beta’s..

    no stutters in the air and the sim works great however at EGLL and some other airport’s I lose 20+ fps and the stutters start… even tho the scenery its self is only showing a few buildings… if I remove the folder for aerosoft egll everything works great and the sim works as it should full settings and clouds, however with the scenery running I cannot run it even it in low settings and NO clouds?…. I submit here and now there is a problem within the airport builds, cars etc.. something with in the scenery is causing me HUGE problems in VR and a major loss in FPS..

    Russ

    Russ

  22. Thanks for making the 2D aircraft shadow visible earlier again! It makes watching landing replays even more fun 🙂

    B17 running well here!

  23. Well Ben, I recognize your work, but … could you end the “stuttering” and “almost exclusive” use of the thread core 1 virtual (in my case -i7 4790K-4.4GHz) at once…?
    451.67 NVidia

    1. XP engine is Legacy and inefficient. It needs a total ehgine rebuild. The core count and graphics card power counts for didley squat.

      Doug.

      1. So is. It is obsolete from the beginning XP 11. New engine, because the stutter does not let you fly decently … You have messed it up in OpenGL too, since beta 16 …

  24. Are we at the stage where we can start using add-ons while testing the Beta?

    Here’s what I have with Beta 17:

    – i7-10700K, Gigabyte z490 Vision G, 32gb Corsair Vengeance 3200ghz XMB Memory, ASUS Turbo RTX 2080 Ti

    – xPilot for Vatsim, Flight Factor B752, Better Pushback, AutoGate, FlyWithLua, Gizmo64, Skymaxx Pro, Real Weather Connector, xsaitekpanels (all latest and greatest versions compatible with Vulcan).

    – Settings: Custom Bezel corrected Nvidia Surround (5012×900) across 3 Acer Monitors.
    Visual Effects: High
    Texture Quality: High
    Antialiasing: 4x SSAA
    Anisotropic Filtering: 8x
    World Objects: High
    Reflection Detail: Minimal
    Vulkan: Enabled

    CPU Usage at Cruise: 7-12%
    GPU Usage at Cruise: 37-87%
    Depending on Clouds

    Everything is playing quite well together!

    I’m getting 45fps looking out the front window of the cockpit.

    However, I am experiencing is a super quick drop to 19.9fps in the F-Sim readout every 5 seconds or so and this also affects the F-Act readout from 45 to 37 for a microsecond.

    This causes small visible “freezes” that are noticeable especially when changing views inside the plane and also when using the chase or circle external views.

    I am almost 100% sure that if I try to go from an external view to an interior view, exactly at the same time as this drastic drop in frame rate, X-Plane will CTD.

    (I was also experiencing the same quick drops with default X-Plane weather).

    I’ve noticed that if I change the Antialiasing to a higher setting, the drop to 19.9 happens more often.

    It also looks like this issue does not occur at all if the frame rate never reaches below 20fps during a brand new flight.

    But if I go through heavy cloud cover on takeoff and/or on approach and the FPS naturally drop below 20fps due to weather, then the “19.9 Drop” cycle starts happening for the rest of the flight, even if the base frame rate is over 40… looks like there’s a trigger point that initiates this behavior.

    I will file a bug after I land at KAUS but I wasn’t sure if I’m just a “cheaty cheater” because I’m using so many add-ons at this stage of the Beta!

    Cheers and thank you for your amazing work!

    Ben

    1. Never removed any, only time you need to remoive addons is when it crashes to see what is causing the crash and so you can tell teh developer. If they dont know they cant fix their adddons.

      1. It’s definitely worth checking whether you have the latest for every add-on – lots of developers jumped on instancing for 11.50 so they could have 3-d objects under Vulkan/Metal, and other developers also fixed long-time threading issues that came to light in the 11.50 beta.

  25. Hi,
    Great work Laminar X plane in VR is a pleasure to fly.
    I don’t get GPU time when I display the frame rate output. I get CPU about .02 but the GPU only shows .001 or similar. Is this a known bug?

  26. Guys, not a bug report but an observation from testing Big Sur Beta 3 on my development machine. I know these are early Apple builds but I thought you’d like to know I noticed significant micro stutters, view change lag and appreciable FPS loss. I can’t see anything in the beta release notes as to what Apple is doing with Metal, so heads up if anyone logs issues on Mac on the Big Sur betas it’s not great at this stage 🙂

  27. Hi Ben,

    I found a timemachine back up of beta 1 and checked performance just in case I was imagining things. Performance has definitely decreased in beta 17 by 10-15 fps in identical situations, settings, addons and custom scenery. Question is are you interested in this as a bug or is that the cost of fixing crashes. I can run the fps test scripts and do like for tests if you wish.

        1. I totally understand that but what I’m asking is if that’s the case, why would I be GPU bound in beta 17 but not betas 1 -14? Should I expect to get the performance back?

          1. Performance of what?

            Beta versions are not usable products, they are inbetween steps on the development path, with (a lot) of bugs, maybe procedures that will not survive the test etc. etc..

            If the final release will be faster then the previous one, it doesn’t care whether or not one step in the developement path was faster but not stable.

            Sorry, it was not me you asked, but it is very disapointing to read all that post about “give me back the previous unstable beta ”

            I’m very happy the way X-Plane develops!

            Wait or go back to 11.41 and enjoy flying.

          2. Hi Günther,

            We are late enough in the beta process that while there may be further perf improvements, it is also likely that for most users the current performance is the final performance. Based on what we’ve seen, I believe that we have _specific_ perf issues, not a general perf problem, e.g. it works well for some users and some users see something dreadful for some specific reason – that’s why we collect performance reports.

          3. Because the earlier versions f.e. had blurry textures or less view range = much less work = much faster

            With early betas i had double fps than with current beta but very blurry textures everywhere

            For me less fps is better then less texture or world objects are just some m away

          4. If you had faster FPS with blurry textures, you can determine if texture res is limiting you – turn the slider way down so everything HAS to be blurry and see if the sim gets fast.

          5. In response to Günther, I’m not asking for a previous unstable beta back. I’m asking if the FPS of beta 17 is going to be more or less what the final release with have. Ben seems to have cleared that up now.

          6. If you got 30 fps in 11.41 and get 35 fps, there’s a fair chance that you get 35 fps at shipping.

            If you got 30 fps in 11.41 and you get 5 fps now, your system is _screwed up_ and you’re not seeing what everyone else sees or what we see in our lab…the thing to do is file a performance bug and we can commission an ETW trace.

          7. Thanks for the speedy reply Ben. Just want to clarify what my first point was in my first post. It was simply reporting the FPS drop from beta 1 to beta 17 and if you were interested in a bug report. I’ve reported performance bugs before but never been asked for an ETW trace.

          8. Right – there has just been an insane number of bug reports, so we can’t even respond to them all in person…that’s why it’s always disheartening to see someone file a bug when they are trying to contact tech support – with bugs we can at least file them by category and get a sense of the size of problems. In the case of ETW reports, we also _specifically_ didn’t commission reports for certain builds where we knew we had problems – it takes time to look at the results of each one and we don’t need to see the 400th of the same bug.

            In the case of b1, I’d say it’s not worth it. b1 is so old and has so many bugs that there is no reason to think that the higher FPS are _earned_. A more useful comparison would be against 11.41 and running GL vs Vulkan. The cases we’re interested in now are mostly massive performance _regressions_.

          9. Sorry Jeff, seems I misunderstood your question a bit.
            But in general I stay to my writing.

          10. Thanks once more Ben, that really clears things up. Don’t suppose I could sneak in a final question? In 11.41 I always founds night lighting ate some FPS. In 11.50 Metal, FPS decreases significantly when the sunlight is bright and increases at sunset and darkness. Is that a bug or just something to do with Metal vs Open GL?

          11. I’m not sure why you would see that difference, but it’s worth noting that the actual rendering settings in 11.41 and 11.50 are not the same. It is not unusual to see a FPS difference from day or night; the shadows and night lighting use very different resources.

        2. I’m also trying to understand the whole fps situation.
          I have a gigabyte h310ma mobo, rtx 2080 super, i7-9700k, 32GB ddr4 2666mhz with 480GB SSD.
          I might be wrong, but I think my pc should be able to handle b17 maxed out with little to no issues(pls do correct me if I’m mistaken). However, like many other users the fps drop never ceases to baffle me. At 4k resolution together with the toliss a321, traffic global, visual effects high(hdr), texture maximum, antialiasing fxaa, anisotropic filtering 1x, world objects max, reflection minimal, scenery shadows checked and vsync on, the fps drops to 23-25 fps at jfk airport with scattered clouds. But then at Amsterdam airport with the exact same setup, I get around 32-36 fps. If I bring the world objects down to high, in both scenarios I get about 10 extra fps. Once I’m above the clouds it’s smooth like cooking grease at 60 fps. I’ve also tried to tweek some 3D performance settings in nvidia controle panel but x plane seems to totally igore it and still do it’s own thing.
          Is there any way to run xplane 11 at 4k with world objects at max and get at least 45 pfs as base?

          1. Right – so I think your situation is typical.
            – You’re trying to max EVERYTHING out.
            – But the domain of “everything” goes well beyond what we shipped in the sim – the gateway artists have cranked up JFK, you’ve installed a third party aircraft, XPTG is adding stuff, etc.
            At some point, your machine could no longer hit your FPS target while doing all of that stuff.

            There’s no guarantee that the machine is even being used efficiently. People have taken plenty of shots at us for not using more cores for rendering (something that is a goal but not for 1150), and I know XPTG does do some multi-core work, but just by having add-ons from third parties there’s a risk that everyone sits on the same part of the computer.

            In your case, the thing to do is to compare FPS at 4k vs 2k – it’s quite possible that the problem isn’t the 4K at all but is CPU side.

            I can tell you one specific gotcha: scenery shadows and water reflections both _amplify_ the amount of geometry being drawn by max 3-d objects. So this is sort of the equivalent of having objects on “super high.” If you find that keeping the 4K but turning objects down gets FPS back, you can keep objects maxed out and back off on the shadows or reflections or both.

            This issue of ‘how do we run on hardware X’ is not an issue on the iOS product – the product is sealed – no enhancements. So we know exactly what things cost and we can tune our own artwork to hit the perf we want. In other words, _we_ set the meaning of “max sliders” so we can just make the max _do less_ to hit FPS targets.

  28. I am running Ubuntu 20.04 LTS, and XPlane 11 beta will not see my Honeycomb yoke. It’s also having a hard time seeing my TWCS throttle. I’m not sure what the problem is here.

      1. The yoke is seen fine by the user and by the joystick calibration tool. This is not a device permissions problem. I do however, see that the device is constantly re-initializing itself in the kernel log file. This seems to be a problem with whatever kernel or udev system has been updated for Ubuntu 20.04. Google the problem, others have it, but I guess not enough XPlane users with honeycomb yokes have updated to Ubuntu 20.04 to get this problem looked at.

      1. I followed one of these guides and something went wrong. Do you have one you know works? By the way, the OS sees it, including the joystick calibration tool in Linux. It is only XPlane that cannot see it.

        1. Never mind, it is not a permissions problem anyway. It is some change in Ubuntu 20.04 kernel or udev system that is causing problems. The yoke IS configurable with the joystick calibration tool in KDE, but apparently XPlane is more stringent with USB devices (as it should be.)

  29. If you’re using X-Camera it could be worth to try to “unable” it and see if it impacts your FPS.
    ***NOT*** saying anything is wrong with X-Camera (at all), but my_settings did that I was losing some FPSs.
    Also moving you vertical view a bit forward or backwards can change maybe 5 FPS or even more, so if you don’t really need an exact “angle” you can play with this a bit and maybe gain a few FPSs.
    (Guess this is not Vulkan version related, just general) – just saying.

  30. I think Beta 17 is a keeper… something was adjusted in the new beta that brought the framerate back to normal (is there such a thing as normal?) and like the early betas.
    No graphic settings changes, but framerate now smooth at 50fr – 60fr, heavier workloads go down to the mid-30’s which is well above normal to me with everything switched on and in place, the better news is that the texture quality (including autogen) stays the same well out from the source, this was a major issue, but now also good. A winner, winner chicken dinner moment! My gut says never update again…

  31. “XPD-10898 Panning delay/stutter.”… It has not been corrected. Continue with stuttering, unless the NVIDIA version 446.14 is installed, both in OpenGL and in Vulkan.

  32. Still the same X-Plane that I used to play since 4 years ago. same random crashes same unstable software same bugs and same bad performance (Just a little improvement in FPS in this beta to be honset). I need my money and my time back. You are wasting our time testing your software and you never gonna fix anything. And pleeeeeease don’t tell me that I shouldn’t write this in the blog and comments because this is the only place we can comment about something like this, and pleeeeeease don’t tell me that this is a beta version because we have the same problem in the stable version too. And pleeeeease don’t tell me to report a but because I already did while I believe that reporting a bug never changes anything with this software.

    1. You should not write this comment in these words. Remarkable though that I as a 2 week new user don’t have any problems at all. Almost 55 FPS in 4k with ortho on zl-18 and some plugins. Maybe you are soooo eager to tweak things that it is over-tweaked. Maybe it is time to move on to F1-2020. With some tweaks you can use your yoke as a steeringwheel ans pedals a pedals

    2. For most people, X-Plane seems to run fine. On Steam, during the last 30 days, 92% of the reviews for X-Plane have been positive (which is better than the vast majority of games achieve on Steam). So maybe your PC/Mac or its configuration is to blame. Or maybe you are just overly critical. In any case, complaining again and again won’t help but only waste your time and annoy everybody. So either use a competing simulator if you think it is more stable (I doubt that) or check if there’s a problem with your computer (maybe overheating, maybe unreliable power supply, maybe buggy graphics driver, etc). Also, with regards to performance, I think I remember Ben Supnik of Laminar Research telling you that you should use a reasonable resolution such as 1080p or 1440p instead of 5k. For example, with X-Plane 11.50 and Vulkan, I get 60-100 FPS on 1440p with medium to high settings on a mid-tier 2017 iMac.

    3. They do not have enough potential to improve your product and will not be able to optimize it with a decent engine. Within that, today is the best today … Unless MFS2020 sends them to hell. But of course, to fly I don’t subscribe to anyone …

    4. Lots of complaining and shouting. Actually you don’t have to try beta versions.
      Maybe Microsoft has the right product for you.

    5. – You shouldn’t write this in the blog and comments
      – This is a beta
      – You should report a bug

      To add…
      – Beta testing is optional. You volunteered your time.
      – You got a good 4 years out of it for $60… I would call that a steal
      – A good majority of people have experienced an improvement, so quite clearly they are doing something right.

      1. for me the sim works 98% perfect thanks to ben and he’s team…. keep in mind they do this for free!!!.. your comments are more opinions, I prefer you keep them to yourself…

        we Love X-plane :)… sure few bugs here and there still to iron out like stutters at egll 😉 but that’s the whole point in beta numbers ( Work in progress )….. Vulkan has been a massive boost for a lot of users including myself.

        FS2020 maybe round the corner but reckon it will take another 12 months + to stabilize some huge bugs iv seen and no chance of VR any time soon… im sticking with xplane 🙂

        russ

  33. I am receiving consistent crashing on b17 – every other beta beforehand has not done so, so I would be interested in finding out what has changed?

    The crash occurs consistently as soon as the cockpit shows, including with default aircraft.

    I have sent the crash report using the GUI which appears.

    LOG:

    0:14:32.789 E/SYS: THREAD FATAL ASSERT: rel_assert Error! Possibly an out of range vector! Attach the debugger and see who called this!
    0:14:32.789 E/SYS: THREAD FATAL ASSERT: !Thread_is_inited() || THREAD_is_main_thread()
    0:14:32.789 E/SYS: THREAD FATAL ASSERT: XPCreateGuiPluginWindow
    0:14:32.789 E/SYS: THREAD FATAL ASSERT: Please report this to Laminar Research.
    0:14:32.789 E/SYS: THREAD FATAL ASSERT: C:\jenkins\design-triggered\source_code\engine\plugins\glue\XPCallbacks.cpp:1769
    –=={This application has crashed}==-

      1. I came here to respond to myself but thank you …

        turns out the rogue plugin was XPUIPC – my local IP address had changed and so different to what I had set up in the XPUIPC.INI file. XPUIPC was attempting to display warning about this on startup (which happens right after the cockpit comes into view after starting X-Plane). Unfortunately whatever method it is using to do this causes beta 17 to crash so they will need to fix it.

        Workaround:
        The fix for this is to simply ensure that the correct local IP address is present in the XPUIPC.INI file in the “~/X-Plane/plugins/XPUIPC” folder.

        X-Plane works fine after doing that.

        1. Is it possible your version of XPUIPC is old? I’ve spoken to a lot of developers over the last six weeks about threading issues, and a lot of pending problems have been fixed.

          If you have the -latest and greatest- XPUIPC and a recipe to induce this crash, please do tell the author. When we see a threading crash report we have no insight into what the user was doing, which means 3rd party devs have to just sort of guess what went wrong with thin hints from us.

          1. XPUIPC is very bad with getting a good version, there are many sites with many different downloads, but no site that looks like the official one.

            I have decided to avoid xpuipc tools because this.

  34. Yeah! I just kicked in late with Beta 17 on a (what you would consider) low-end hardware. Starting with about 30 FPS on a clear sky with OpenGL dropping to 24 FPS with clouds, I switched to Vulkan and I got 60FPS for clear sky and about 34 FPS with clouds!
    I could not see any graphical artefacts while flying a short tour.
    The only disadvantage I noticed: My GPU blower is blowing at full speed all the time, and even after terminating X-Plane, it blows like another minute until it slows down gently…
    I was also surprised by the loading times, because I almost felt Vulkan is loading a bit faster than OpenGL. Still it takes quite some time, but not as much as I had feared…

    1. Re: the GPU blower — there are two bugs:
      1. My GPU is idle – why aren’t you using it
      2. My GPU fan is going nuts – it annoys me.

      We can’t fix both. 🙂 🙂

      1. My PC is screaming when I play any demanding game.
        “Demanding” in this case is good, as it means it is actually doing some work 🙂
        @Ulrich – no offense or anything, but maybe check how your filters and coolers are doing?
        I got a pretty powerful PC, but I still need to clean filters, fans etc now and then.
        The difference can quickly be 5-10 Celsius of difference for the CPU, which again can mean a heavy extra load on the cooling.

        P.S. Stop offending each other with F1 “fun” and fanboy-ism, it only brings more sh** 🙂

        @Ahmad Abou Hashem: where are you at the spec demands?
        And have you ( as well) ever tried to give you PC a good cleanup?
        With Vulkan, X-plane has just gone more demanding than ever on cooling if you’re on spec “edge” because it simply “gives you more workload with the same hardware”.
        Stay cool everybody :–)

      2. No, the new “other sim” _surely_ can keep the GPU on 100% load while keeping cold!
        ; )

          1. So or so, it _surely_ will do everything completely multi threaded and all cores will be utilised totally equal all the time. ; )

          2. It’s funny you mention that, because 11.50 is the first version that intentionally categorizes cores into real-time/for the frame and background/loading. We have to make sure that a big background loading task (like a DSF load) doesn’t cause the FPS rate to briefly plummet (by taking away computation resources) – so our solution is just to outright partition the machine into flying and loading. The minimum config that works sanely is a 4-core machine (with 1 worker helping the rendering thread and two doing loading); with more cores we can get wider on both sides.

            But no matter how much concurrency we can add, we don’t want to reach a point where on an 8 core machine spends 8 cores rendering a frame, because the first time you hit a new autogen area or DSF block, frame rate will tank.

  35. Thanks for B17 , we can see final version comming soon.
    I do have a question regaarding you main thread of xplane. Is there a particular reason why this thread is running at 100% cpu (core) usage as if it where looping endlessly ? Even on the main menu without any simulation running and limitting fps of the main menu to something quite low, there is still that main thread of xplane taking 100% CPU of 1 core. I think this thread need some “sleep / timer”.

  36. Hi everyone,

    when i fly x-plane 11 in windowed mode I get 60 fps in some areas with vsync enabled or sometimes 50 or 45 (that’s okay). But when I fly in fullscreen vsync sets the fps at 30 as far as the fps are below 60. Why is that and is that normal?

    Ryzen 5 2600 ( 20 % load in x-plane 11)
    RX 5700 (30 – 40 % load in x-plane 11)
    16 GB 3966 
    Win10 2004

    1. I get the impression vsync only works if XP can use the full screen. So that explains the behaviour you (and I) see.

  37. Just thought I mention. I am experiencing major FPS drop when I have Firefox open in the background. It seems like it is using a little GPU, (not much though, only 0-2% GPU according to the task manager), but enough to have a very noticeable impact on FPS and stuttering in X-Plane with Vulcan.

    1. That’s beyond the scope of what we consider when dealing with stutters. It’s hard enough guaranteeing stutter free operation on a PC, it’s harder still unknown plugins and tweaks are running, but once you say “and my PC is also doing X”…

      1. Stuttering is higher outside of NVIDIA 446.14 on Vulkan. You will have to invent another excuse because it is not valid (you have too many). And it continues to abuse the second core (100% virtual thread CPU2) to the point of heating the processor to the specification limit, even with very good cooling of the case, the mobo, … everything and … NOR THOSE. IT’S VERY BAD OPTIMIZED FOR CPU (DOES NOT USE ALL THREADS), whatever. Since you do not fix that, I see that you are not capable, you are going to lose many clients.

        1. Dude, if having a single CPU core at 100% is overheating your CPU, then your cooling is broken (definitely not “very good”) and you should go get it fixed. You might also want to consider that when the CPU gets throttled from being too hot, that would be the source of your stuttering

          And if having one CPU core doing work is cooking your computer, I have no idea why you’d want more cores doing stuff…

          1. I didn’t say it overheated. It reaches maximum points 71ºc for being used, mainly, a single core and it is happening in more than one processor of all kinds. With 38ºc mobo and harward and Windows 10 optimized so that all threads are used CONSTANTLY LESS 1 (for Windows) with air cooling and extraordinarily ventilated all in the most optimal way. This software has always been like this and Ben has already said that it is not his priority … The “stuttering” thing is embarrassing and the “balls licking”.

      2. Hello Ben,

        I did not actually consider this to be an issue with X-plane or anything I expected you to fix. It was more just for it to be mentioned somewhere, as I don’t remember ever seeing it written. I am sure most people would not consider just having a web browser with SkyVector and a few other, similar tabs open in the background as a source of performance issues.

        There is a similar loss of performance in Half-Life Alyx when I have Firefox open in the background. It then often complains of low VRAM even though I have a 2080 TI, and noticeably scales down the visual fidelity(however, without FPS drop, but I know they have some fancy logic to ensure this is the result). So this is definitely not an X-plane issue.

  38. Thanks for the reply, it seem you r are right about VSYNC ON, prevent the high usage of the thread. I was at first limiting the FPS with MSI AfterBurner but it wasnt calming the thread.
    But my main concern is still there, in menu or in-game im suprise how easilly this thread can go to 100% for nothing. Ex: looking at the sky with Vsync, with the simulation paused, I get hardly 120fps and the main thread is super high in it’s usage, I think this is anormal and need optimization. Just my 2 cents.

    1. This just isn’t first on my priority list. Like, if the bug is “I use x-plane, one of my cores is fully used and I have 120 fps” … I must worry about other things before this.

    2. Are you running Windows? I think you are. On Mac and Linux, 100% utilization is a desired state. They start having to sweat at that point. Windows gets “angry” near 100%

      Not sure why.

  39. 1. Don’t forget to send Microsoft a cake on their release of MSFS 2020.

    Cake ‘could’ include…..

    a) MSFS logo with the Windows logo
    b) XPlane logo with Windows, MacOS and Linux logos.

    regards

  40. Loving B17, trying it out on my laptop ( ASUS ROG Zephyrus 16G RAM, RTX 2080 8 G ) and an HTC Vive. There there are some surprising interactions and effects across:

    1. Nvidia Global settings
    2. Nvidia settings for Xplane ( threaded opt. OFF)
    3. Xplane settings and datarefs ( balloons, forests, LOD )
    4. Settings for VR ( native eye res, oversampled, smoothing etc.. )
    5. Adjustments for (x) rig drivers, addons, etc.

    that drastically affect the fps/stutters ( actually observed ) in the cockpit.

    Speaking purely for myself, maybe collecting the overall wisdom of this community in coming up with a strategy ( starting with a vanilla install ) to get the best “possible” performance from your rig might be a good thing… time we had an updated guide out to help out new virtual pilots.

    The new “other” sim looks awesome, but dont forget, x-plane has had ( and improves ) flight dynamics, flight systems, weather, realistic terrain, osm scenery, ortho scenery, multiplayer, real traffic hooks, fmod sound, pbr rendering, multi-machine, multi-monitor ….new UX and native VR support! Amazing work from the LR team!!!

    We users will never be happy until xplane has missions, a “proper atc”, awesome clouds, and photoreal scenery, eaves, and a full authoring and editing toolset built into the product ( wait, scratch that one )…. blender support? 🙂

    1. I think there are soooo many permutations (just follow the forums on x-plane.org) that it is next to impossible to create a guide for this

      1. Hmmm … i disagree… talking about a strategy.

        Assuming your PC is running at max speed, fresh boot into windows, single monitor, nvidia setup for performance, everything off in the nvidia settings, xplane at hdr, fxaa, aniso off, medium world, no reflections/shadows,vulcan, running -tutorial , no addons. 16 G RAM, 8 G card, i7 with 8-12 cores. What max fps should one expect?

        And where do you go from there to go faster without dumbing down anything?

        2nd , if its all downhill from there at least you know what your system baseline is, any improvement without sacrificing the experience is gravy.

      2. Threaded Optimization should have no effect for vulkan, its a setting of the Nvidia OpenGL driver. Maybe Ben or Sid can enlighten us about this performance myth.

        1. Mostly true – except that _plugins_ could be affected. My expectation is that given the complexity of our GL/Vulkan bridge, threaded optimization would make GL plugin drawing worse. But we haven’t tested it – our view is that users should keep it off and plugins should write GL code that doesn’t spend a ton of time in glBegin/glEnd.

          1. WE recommended off for pre-11.50 too. Since threaded driver launches a bunch of background threads and so do we, at times when everyone chooses to run at once (DFS load, autogen, water, AI aircraft) we would pre-empt the NV threads and the driver would die in a spin lock.

  41. Complete fresh installation of XP. Running B17. My GPU is only working at 20% CPU is working at 60% thereabouts. All plugins removed complete vanilla. Panning the view in 2D I drop 15/20 fps momentarily, this recovers quickly so not an issue 2D. However, I am a VR flyer and I experience the same in VR too, this means FPS drops below 27 to 15 then recovers to 26fps and eventually stabilizes at 27, until I breath! Same with aiming for 40fps. Performance for me has drastically dropped since B14. Also I notice that weather chews up around 55fps – is this normal? I have filled a bug for the above just looking for an answer. B14 for me was like being given a Ferrari to drive for a month then someone took it off me and told me to drive a Prius 🙁 especially where VR is concerned.

  42. Hi All

    Processor Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz, 2592 Mhz, 4 Core(s), 8 Logical Processor(s)
    GPU – NVIDIA Geforce GTX 970M
    16GB Ram

    On previous versions I could fly in all weathers with an average FPS of 23 with this version I can manage 32 FPS with clear weather but still not smooth when I pan etc, with clouds on I can only get 10 to 15 FPS max.

    Any ideas? I have tried updating GPU Drivers and a fresh install of X-plane 11 but still no luck.

    Thanks

    John

  43. Thank you guys for fixing the nasty water reflection artifacts in VR! You really only appreciate how bad they were when they’re gone. 🙂

  44. good evening impossible to run the vr with rifts. the screen remains black. how to revert to beta16 or everything was correct. In addition, a window appears asking me to buy a “time expired demo” pro usb key

  45. Any advice what to do?

    I have just switched from 11.41 to 11.50b17. During my two flights I made on 11.50b17, I experienced several FPS freezes. I can’t specify the root cause of the freezes as the appeared in various situations. Once while being on ground, the other time while moving from external view to cockpit view… There was no CTD, only sudden freeze and then FPS went back to normal.

    (Un)fortunately log.txt is clean, no indications of something going wrong. And at this moment I’m running out of options. Shall I do remove-all-plugins-and-add-them-back-one-by-one kind of investigation or there is some other, more time efficient method of looking what slows down X-Plane?

    I have i9-9900 with 64GB RAM and NVidia 1080. During the FPS freeze both RAM, CPU and GPU loads aren’t maxed out. CPU and GPU are utilized around 50%, as per Windows’ Task Manager.

    Help! 🙂

  46. Get going..

    I am confident: X-Plane is going to be the main competitor.

    So start confessing to it.

    There is no (0) point in riding it out.

    1. Fully Agreed,,,

      By the way, based on released betas of M$FS and influx of many public videos recently, i would say they suffer from a LOT of STUTTERS.

      While X-Plane is putting STUTTERS as No. 1 Enemy,,,
      So by time Laminar smooths out these rough edges, they can unleash efforts to beautify x-plane with new next-gen environment

      Technically, the distance or (GAP i would say) between XP and FS2020 is not much
      XP Basically needs:
      1) Overhaul for all textures of planes inside-out and scenery
      2) Weather System replacement supporting smooth multi layered volumetric clouds
      3) Better AI Ground Services and probably 3d duty-free shopping corner in airport
      4) better 3d trees
      5) enhanced airport mirage effects (cinematic graphics effects)
      6) more accurate 3d models for pilots and ground services personnel
      7) better daylight sun glow and netter night lighting
      etc…

  47. With reference to XPD-10895 ACF shadow cut off when progressing time:
    I still get aircraft shadow cut off’s when progressing the time. It happens early in the morning and late in the evening. Should I file another bug for this?

    1. Yes! We list the bugs as fixed so that if you still see it we can get a new report – please include as much info as you can about the conditions in which it happens (rendering settings, location, etc.)

  48. Hello Ben,

    It is not probably the best place to discuss about that but have you tried x plane 11.50 on an apple DTK platform (ie Mac mini arm)? Will 10.50 also compiled for this platform or run with Rosetta 2? what are the perf to expect?
    My current iMac 2013 need to be refresh and I m hesitating to take the last 2020 intel or postpone to next arm.
    Can you please give us your vision and roadmap ?
    I think this will interest a lot of people here.
    Thank you.

    Kind regards,

    André

    1. We are under NDA with Apple so there -is no- appropriate place to discuss DTKs with us. 🙁

      I can say a few things:
      1. Someday there will be a build of X-Plane that officially supports ARM hardware and whatever OS it ships on – this will be our 5th ABI port and we already compile the iOS product for Apple ARM chips, so we don’t expect this to be problematic.
      2. There will definitely _not_ be any back-porting of existing end of lifed products to the new ABI. So absolutely no way there’s an ARM-native v10. There is no guarantee that 11 will support ARM – it doesn’t now and we have not announced support yet – just that we will do it at some point.
      3. For what it’s worth, I just ordered the new August 5k iMac today. My situation may not apply to you though – I have three dev machines, all paid for by LR, and two of them are Apple hw, so I can easily run one of each for ARM/Intel, and there’s no cost. Since my current dev iMac is from 2014, I was _really desparate. (The fact that ordering the new machine and paying was faster than compiling the terrain shader on my iMac is a sign I needed the upgrade.)
      4. We expect the Apple ARM chips to be very good! The iPhone X and 11 just eat code for breakfast. So based on benchmarks and experience we are not at all worried that the new ARM chips would have any perf problem.

        1. Plugins compiled or x86 will work if the sim is running in x86 mode. Plugins compiled for ARM will work if the sim is running in ARM mode. But you can’t cross the two.

          When the sim ships with ARM support it will be one app with both ABIs supported, so it “just works” everywhere; at that point it will try to load ARM plugins _unless_ you check the box to force x86 mode in the Finder. (All of this infrastructure is exactly like the 32->64 and PPC->x86 transitions, because the dual-ABI executable file format Apple uses has been around since the x86 move.). So if you have an ARM mac you’ll be able to run in ARM mode when (1) the sim supports it and (2) all of the plugins you need have been recompiled with ARM as well as x86.

          1. Thanks Ben for sharing this.
            You didn’t mention Rosetta 2, does it means that we won’t be able to emulate x plane x86 on arm ?

          2. Oh uh…yeah. We expect that you’d be able to run older versions of X-Plane on an ARM mac under emulation. But we don’t know how well that will work and if it doesn’t, it’s not something LR would address (e.g. if the bug report is “v10 doens’t run on an ARM mac in emulation on Big Sur”, we’d be like “it wasn’t supposed to.” So we’ll have to see how it works when the tech is available.

  49. I wonder if it would make sense to build 2 seperate binaries for OpenGL and Vulkan instead of running permanently through “if (vulkan) then doVulkanStuff; else doOpenGLStuff;” decisions. Code would probably be simpler and faster. The source could be still in one piece and conditional compiling and linking just produces a leaner executable. Just my 2 ct

    1. I don’t think it would make a difference. Since our graphics abstraction is modern, we don’t have cases where we have to make a large number of small function calls to accomplish work, so the cost of the dynamic dispatch is quite small. If we did have a ton of HAL function calls to do stuff, we’d already be writing bad graphics code.

      The run-time abstract interface dispatching gives us better code structuring than conditional compile I think.

    1. That implementation of 3-d water never shipped because it had unsolved problems in _real_ test cases – it was a cool tech demo. We _do_ want to make water displacement real 3-d some day, and I expect to make my _third_ attempt at such a feature at some point.

Comments are closed.