Announcement

Collapse
No announcement yet.

Distributed Bucket Priorities

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

  • Distributed Bucket Priorities

    It would be super-useful and amazing to be able to change priorities for distributed renders after they've been sent. The UI for it could be adding a simple priority spinner right inside the render view.

    Problem: I am rendering a print render on 20 machines and it is taking longer than expected. Another rush-job comes in to another artists and we have to cancel the render to free the farm to start the the rush job.

    Solution: Fire off the new job even though one is already rendering. Make the priority higher in the new jobs render window and lower it on the one that is already going. Then it can finish what buckets are going, and start loading the next scene without missing a beat. Then one the rush is done, the other render will pick up right where it left off without having to cancel, or do region render resumes.

    Bang... awesome.
    -Joel E
    https://www.biglittlepictures.com

  • #2
    Yes, however it's easier said than done In general we've been trying to keep away from writing a network job manager, as that opens a whole new can of worms - we can easily spend years just on that and neglect to work on V-Ray itself. I know that Deadline has some tools for exactly this purpose, perhaps it might be useful to look into it.

    Best regards,
    Vlado
    I only act like I know everything, Rogers.

    Comment


    • #3
      I can completely understand that. We use Smedge to manage network renders of many users and animation frames. Print renders kind of behave differently though because of the length of time on a single frame and how they can clog a farm if something shorter term is needed. Perhaps something as simple as the ability to uncheck and check active network computers AFTER hitting render, somehow accessing the settings dialogue. This would also give us the ability to manage our farm after a render has been fired without stopping it, and shouldn't be out of scope. This would solve our problem

      I can totally appreciate your sense of focus on core VRay, this is why it is the best off the shelf product on the market. You guys do great work.
      -Joel E
      https://www.biglittlepictures.com

      Comment


      • #4
        If the window that pops up in the Render Setup/Settings/Distributed Render is accessible in maxscript, we should just be able to call the command and tweak some parameters in the existing UI. It seems like you can type maxscript even while rendering. Does that particular window have a maxscript command call?
        -Joel E
        https://www.biglittlepictures.com

        Comment


        • #5
          Originally posted by joelly3d View Post
          It seems like you can type maxscript even while rendering.
          Hmm, can you? How do you do that? If it is possible we can certainly add MaxScript functions for this.

          Best regards,
          Vlado
          I only act like I know everything, Rogers.

          Comment


          • #6
            Seems like typing in the mini-listener works. I was also thinking, we could edit vray_dr.cfg manually if we had to, but after you edit it it doesn't reflect the updates update automatically. Probably because max doesn't continue to read that file constantly but only on render call.... i assume.

            Either way, if we could force a re-read of the file either with a GUI pop-up, or a refresh the load of the cfg, it should get us something that could manage what we need as a short term solution. Seems like making the Distributed Render Settings window available in the renderview or VFB window would be the best and easiest from an end user perspective ultimately.

            Thanks for looking into this. It would be a massive help.
            -Joel E
            https://www.biglittlepictures.com

            Comment

            Working...
            X