Announcement

Collapse
No announcement yet.

RT GPU DoF

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Thanks Vladimir, DoF and Cuda all working well thanks.

    Just a general question, launching a render in standalone always takes longer than rendering with RT GPU within Modo.
    I imagine this is not surprising, if it's loading everything onto the GPU.

    Are there any suggested ways to optimise? Rendering an animation for example seems to take time loading scene per frame...

    Best,

    Nick

    Comment


    • #17
      It takes more time because a .vrscene file with the whole animation needs to be written out to disk.
      And then read back by Standalone on each frame.

      If you have non-animated Replicators you can mark them as such, which will make the .vrscene smaller for an animation.

      Also, you can enable V-Ray subdivision on your meshes. This will cause only the mesh cage to be written to the .vrscene.
      If you have deforming meshes in an animation, this can save a lot of time. Note that the V-Ray Subdivision might give you slightly different results.

      Greetings,
      Vladimir Nedev
      Vantage developer, e-mail: vladimir.nedev@chaos.com , for licensing problems please contact : chaos.com/help

      Comment


      • #18
        Thanks Vladimir, the DOF working good again on GPU on Mac Os!

        Comment


        • #19
          Hi,

          Is this loading from disk then, as opposed to the scene held in memory within Modo?
          If so I guess a faster SSD would help speed up the loading time?

          Originally posted by vladimir.nedev View Post
          It takes more time because a .vrscene file with the whole animation needs to be written out to disk.
          And then read back by Standalone on each frame.

          If you have non-animated Replicators you can mark them as such, which will make the .vrscene smaller for an animation.

          Also, you can enable V-Ray subdivision on your meshes. This will cause only the mesh cage to be written to the .vrscene.
          If you have deforming meshes in an animation, this can save a lot of time. Note that the V-Ray Subdivision might give you slightly different results.

          Greetings,
          Vladimir Nedev

          Comment


          • #20
            Originally posted by nickjamesgordon View Post
            Hi,

            Is this loading from disk then, as opposed to the scene held in memory within Modo?
            If so I guess a faster SSD would help speed up the loading time?

            Maybe, can't say for sure.
            Is it really slow ? How big is the .vrscene that gets exported ?

            Greetings,
            Vladimir Nedev
            Vantage developer, e-mail: vladimir.nedev@chaos.com , for licensing problems please contact : chaos.com/help

            Comment


            • #21
              It's not slow! But rendering animations, so trying to get each frame done as quickly as possible.
              I saw someone mention 'split-rendering' in a facebook group.

              Could I fire up several V-ray standalone renders, for each GPU on one machine?

              Best,

              Nick

              Comment


              • #22
                Yes, just modify the CUDA devices before starting the render.

                Greetings,
                Vladimir Nedev
                Vantage developer, e-mail: vladimir.nedev@chaos.com , for licensing problems please contact : chaos.com/help

                Comment

                Working...
                X