EXPERIMENTAL USE ONLY!!! -- [What?] -- Login to add public annotations [How? >>YouTube].

As you may have heard, we’ll be at FlightSimExpo 2019 in Orlando this weekend! If you are attending, please stop by our booth and drop in on our talk Saturday afternoon. We’ll be qualifying people for a VR landing contest too.

For developers: Austin, Philipp, Chris and I will all be there – find us to talk about developing add-ons for X-Plane!

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.

28 comments on “We’re Going to Orlando”

  1. Video this time? we all can’t get to Orlando, so this is our only way to see and join the event… Vegas was a mess? and not very professional…. have a great show.

    1. Hopefully _live_. The guys who were supposed to do it last year flaked out, but this year my understanding is that the event organizers hired a professional and I think there might be a live stream.

    2. FSElite have a hub for the expo on their site and they will be live streaming all of the presentations along with interviews with devs, including Laminar 🙂

  2. Are you going to talk about perfomance boost numbers in X-Plane with Vulkan compared to OpenGL?

  4. Hi, will we see anything about Vulkan? Those performance tests you talked about some time ago?

    1. It will be a subscription option in the current app. No features of the current app are being taken away – if you don’t use it, you still have access to everything already there.

  7. Great show. 11.35 sounds like a fantastic patch. Also great to hear there is light at the end of the tunnel with Vulkan but I note you focused on potential GPU gains and didn’t mention the possibility of CPU gains by utilising extra threads? Given most people are CPU bottlenecked I was hoping to hear something on that.
    Congrats to Austin on his new baby 😉

    1. When you talk about “potential GPU gains”, I think you’re referring to “things we can do later after Vulkan once we’re using Vulkan and part of this new ecosystem, right?”

      If so, yes, the gains are not just in improving GPU performance via insight into the low level behavior of the GPU with better tools, but also the ability to multicore the engine since we don’t have threading driver restrictions. I just didn’t have time in the slides for it. But Vulkan makes multicore possible.

      1. I meant that you talked of the “outofthebox” gains you are getting from Nvidia and AMD GPU’s but didn’t say anything on potential CPU gains so people who watched may be a little misled and/or disappointed in the idea that those were the gains for Vulkan period.
        You have explained that you had slides but didn’t have the time and that is understandable, you are spinning a lot of plates and you will be gathering data after Vulkan beta release to see how CPU usage can also be best optimised so don’t want to make bold claims.
        Perhaps you could show those slides and maybe share some thoughts in another blog down the road?
        My enthusiasm for vulkan is because I can see the potential of Xplane but currently the engine is running on mud and even high end systems struggle with frames if we want to push the graphic capabilities and enjoy it to anywhere near the full and given I am a VR enthusiast that problem is exacerbated even further.
        I have a pimax 5K+ and the potential is obvious, unfortunately so many sacrifices have to be made to keep up performance though that the actual experience is far removed from what it could be – which is jaw dropping.
        Sorry for the long post.
        Rooting for you guys and Vulkan.

        1. I’ll answer for Ben here! It’s actually the other way around, the gains Ben showed in the presentation are all on the CPU side. Vulkan is mostly aimed at reducing overhead from the CPU side coming from the OpenGL driver. The performance gains on AMD and Nvidia systems aren’t because X-Plane utilizes the GPU better, but rather because we no longer rely on the AMD and Nvidia OpenGL driver to do the right thing ™.

          But as Ben said, this is only the beginning. We can do more tricks now such as multithreaded command generation, to further decrease CPU load. But we also get access to GPU profiling tools, so even though it’s initially only aimed at CPU cost reduction, we can also look at what the GPU is doing and improve things there as well.

          2. okay so now I am thoroughly confused.
            I thought what was shown was gains on Nvidia and AMD GPU drivers running Vulkan.
            I just checked the stream and it may be because the first 2 performance slides show graphs for Nvidia and AMD which obviously we associate with GPU not CPU and then in the next graph you show RX480 and GTX980TI gains again GP{U’s not CPU’s.
            If you were talking about CPU gains I would of thought you would have a graph that showed something like Intel i7 2700 before and after vulkan or something?
            Obviously I am misunderstanding how the data is being presented but the video is below for reference.
            Thank you very much for trying to explain it to me though 😀
            https://youtu.be/WuUaq7_9CLo?t=17199

          3. Sorry for the confusion. The graphs are indeed showing an AMD GPU vs an Nvidia GPU, however, the performance improvements are all on the CPU side! The FPS in the OpenGL graphs is limited by the CPU side in both cases, and a part of the limitation is because the OpenGL driver has a tremendous amount of overhead and is on the CPU side. Keep in mind, the OpenGL driver is implemented by the hardware vendors themselves, so AMD and Nvidia have their own implementation. And with Vulkan, that implementation is what we are sidestepping and why we see improved CPU utilization. This is also why the results are different with AMD and Nvidia, the quality of their drivers are different.

          4. Thank You Sidney.
            You have all been very patient and gracious with your replies so need to respond again, I’m just thinking aloud.
            I think I have finally grasped what was being presented.
            CPU gains through OpenGL calls on Nvidia and AMD GPU’s utilising Vulkan API?
            I would expect the gains to be different depending on the CPU/GPU configuration as well though – how bottlenecked the CPU is and therefore different results (gain) based on different graphic settings within Xplane but you were just giving a middle of the road example in your presentation.
            I bet you are sick to death of Vulkan so thank you again.

      2. Great presentation, looking forward to the Vulkan release. On another topic, with the success of the Scenery Gateway, ever thought to extend it to Cities? i.e. define polygon city boundaries, give them an Id, and let developers create and upload them to a City Gateway?

        1. We still are planning to use autogen for cities – we think we can make significant improvements in the translation of OSM data to X-Plane.

          2. That sounds very promising for VFR and those who love eye
            candy if I am reading between the lines correctly.

  9. Wasnt surprised about ATC. but a bit dissapoined with comments about , dont you all Fly VATSIM. Im no professional and my short attention span would preclude me from flying online, after 15 years of simming im only just beginning to do procedures. I think like many would simply want a ATC that gets you from airport to airport without flying you into mountains and having you change course 200 times to get lined up. All the rest can come later, I normally dont comment but this is the only let down i have in xplane, which I am having more fun than I ever did, so much more immersive.

    1. I was disappointed with the response as well.

      In early 2018, 240 comments were submitted to this blog on improvements users wanted. The X-Plane.org forums are filled with users baffled by the lack of functionality. I see at least a couple new posts on the ATC a week!

      I recently read a 2011 blog post by Chris on the X-Plane 10 ATC system. I stumbled across the post while figuring out how to build my own ATC TRACONS from data I got from the FAA via a FOIA. I was happy to see the new voice system, thinking that might bring some of the features discussed in that blog post 8 years ago, but I’m not so sure now.

Comments are closed.