Page 35 of 50 FirstFirst ... 25333435363745 ... LastLast
Results 511 to 525 of 744

Thread: Octane render for Lightwave 2.0

  1. #511
    Quote Originally Posted by HenrikSkoglund View Post
    Oh, that's too bad, good that Octane seems to go gently on the memory then ��
    Well it's bad and good in same time. Since all GPUs have same info/data and work in parallel you can stop any GPU in middle of renderign and you will loose only those samples of that particular GPU but rest will continue smoothly and then you can re-add GPU(s) back in middle of renderign and it'll work fine and just be faster . In case memory would all be combined than that wouldnt wok but since it's all paralel you can do that easily. Any no flickering or GI baking issues .
    RAM-Studio
    WS - Dual Xeon E5-2698v4/128GB/Win10x64/x TitanX(M) + 2x GTX 1080Ti + GTX 1080
    My LWM Video Car Modeling Tutorial

  2. #512
    Super Member spherical's Avatar
    Join Date
    Dec 2004
    Location
    San Juan Island
    Posts
    4,686
    Quote Originally Posted by HenrikSkoglund View Post
    * Do you need as much memory on the graphics cards as system memory?
    Not really practical/possible. Most systems would bog down if they were limited to the largest amount of RAM that is available on a graphics card. IOW, for a system to run efficiently, it needs 16GB-32GB RAM. Maximum graphics card VRAM, at this writing, is 6GB-8GB.

    Quote Originally Posted by HenrikSkoglund View Post
    When I package the scene as a .zip file in Lightwave it's about 2GB flat.
    A zipped package is irrelevant. First, it's compressed, so that's not telling you anything of real use. You need to check how much RAM is being used when the scene is loaded. An easy way to keep track of this is the Render Status window. Hit F9 and on the right is the amount of memory in use for that render.

    This doesn't mean that an external renderer like Octane or Thea will require the same amount on the graphics card, as they are different technologies and handle assets in other ways so as to fit heavy scenes into available VRAM.
    Blown Glass · Carbon Fiber + Imagination

    Spherical Magic | We Build Cool Stuff!

    "When a man loves cats, I am his friend and comrade, without further introduction." - Mark Twain

  3. #513
    Quote Originally Posted by spherical View Post
    Not really practical/possible. Most systems would bog down if they were limited to the largest amount of RAM that is available on a graphics card. IOW, for a system to run efficiently, it needs 16GB-32GB RAM. Maximum graphics card VRAM, at this writing, is 6GB-8GB.



    A zipped package is irrelevant. First, it's compressed, so that's not telling you anything of real use. You need to check how much RAM is being used when the scene is loaded. An easy way to keep track of this is the Render Status window. Hit F9 and on the right is the amount of memory in use for that render.

    This doesn't mean that an external renderer like Octane or Thea will require the same amount on the graphics card, as they are different technologies and handle assets in other ways so as to fit heavy scenes into available VRAM.
    Hmm, yeah good points. This particular scene is taking up around 18 gb ram memory and would probably not fit inside any of these gpu's limits then. But it felt kind of like an eyeopener for me since it quickly gets expensive to put work like this on a renderfarm. I needed to get 1500 HD frames out of this one and the client asks "can I have them tomorrow?".

    Gosh...

    Anyway, I was thinking maybe one could split the scene into parts and render with mattes if one were to use Octane?
    Also I read something about memory unification in cuda 6 combined with newer gpu cards, but maybe it's for some other use.

  4. #514
    Super Member spherical's Avatar
    Join Date
    Dec 2004
    Location
    San Juan Island
    Posts
    4,686
    Most of the external renderers do bucket rendering now, so they may be able to handle something that big. Unified memory is coming but haven't seen anything available on the market as yet.
    Blown Glass · Carbon Fiber + Imagination

    Spherical Magic | We Build Cool Stuff!

    "When a man loves cats, I am his friend and comrade, without further introduction." - Mark Twain

  5. #515
    Lightwave rebel Pensart's Avatar
    Join Date
    Feb 2003
    Location
    Belgium
    Posts
    291
    Playing with the octane demo version. When using octane object properties - subdivision sharpening, is there any way we can assign which edges should have sharpening? In modeler i have assigned 100% cc sharpness to some edges but octane sharpens all the edges of the object.
    -------------------------------------
    Joined the Rebellion

  6. #516
    Nope, currently the OpenSUBD surfaces in the octane plugin don't support edges sharpening, but remember that you can render the LightWave native Subpatch or CC SDS surfaces without problem, like in the LW native renderer, if you need this feature.

    -Juanjo
    Juanjo Gonzalez
    Computer Graphics Software Developer

  7. #517
    Registered User
    Join Date
    Jul 2007
    Location
    Belgium
    Posts
    114
    Hi, when will Octane's out of the core rendering feature be available? It's on their roadmap.

  8. #518
    In theory the support of the out of core textures should be available along this 2.x dev cycle.

    -Juanjo
    Juanjo Gonzalez
    Computer Graphics Software Developer

  9. #519
    The second beta of the new OctaneRender™ 2.2 for LightWave™ plugin has been released:
    http://render.otoy.com/forum/viewtopic.php?f=36&t=44731



    -Juanjo
    Juanjo Gonzalez
    Computer Graphics Software Developer

  10. #520
    Nice implementation of light passes!

  11. #521
    Registered User
    Join Date
    Jul 2007
    Location
    Belgium
    Posts
    114
    Off Topic, but I'll ask it anyway.

    Juanjgon, could you develop a plugin the use of 3Delight in Lightwave? It's not as good in terms of image quality as Octane render or Lightwave's native render, but it's very fast.
    For small teams with 2 or 3 computers you can render a whole lot of animation. It's very fast in displacement and motion blur, but not very good in raytracing. I've experimented with it and like the speed of it. If FiberFx and hair would also be supported, that would make my day. I would spend 100 to 200 USD for a plugin on it.

  12. #522
    Quote Originally Posted by Curly_01 View Post
    Off Topic, but I'll ask it anyway.

    Juanjgon, could you develop a plugin the use of 3Delight in Lightwave? It's not as good in terms of image quality as Octane render or Lightwave's native render, but it's very fast.
    For small teams with 2 or 3 computers you can render a whole lot of animation. It's very fast in displacement and motion blur, but not very good in raytracing. I've experimented with it and like the speed of it. If FiberFx and hair would also be supported, that would make my day. I would spend 100 to 200 USD for a plugin on it.
    As you can imagine, develop a complete plugin to support a external render engine is really a complex task. Currently I am very busy with my projects, so at least this year I am not going to have free time to begin a new one

    -Juanjo
    Juanjo Gonzalez
    Computer Graphics Software Developer

  13. #523
    Hi Juan, I have just tested the latest demo today and it was such a big improvement compared to my last try with 1.2. Thanks a lot for bringing Octane to LW. I'm running this on a single Titan Black, performance was pretty good way better than my expectation. I'm seriously considering in purchasing a license very soon.

    I have a small request, could you implement Ctrl+Alt+drag in the IPR camera mode to zoom and Alt+drag to pan? The current mouse wheeling zoom isn't convenient for us tablet users, having to reach for the mousewheel to zoom in and out is just unintuitive. Viewport navigation like Fprime is much preferred. Would be great if we could also implement shift+clicking in IPR to surface pick.

  14. #524
    Super Member gordonrobb's Avatar
    Join Date
    Oct 2003
    Location
    Scotland
    Posts
    1,323
    Fairly sure shift-click does already pick the surface.

  15. #525
    Quote Originally Posted by gordonrobb View Post
    Fairly sure shift-click does already pick the surface.
    You are right, it does. Strange that in my initial test, I couldn't get it to open surface and node editor with shift-click. Now it does.

    It could be I was trying to pick it while IPR is paused, which it wouldn't.

Page 35 of 50 FirstFirst ... 25333435363745 ... LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •