Jump to content


  • Content count

  • Donations

    0.00 CAD 
  • Joined

  • Last visited

  • Days Won


malexander last won the day on June 19

malexander had the most liked content!

Community Reputation

346 Excellent

1 Follower

About malexander

  • Rank
    Houdini Master
  • Birthday 03/06/1974

Contact Methods

  • Website URL

Personal Information

  • Name
    Mark Alexander
  1. My RTX 2080 experience

    Do you have a power supply that's in the 800W+ range? Usually if things are shutting down without warning it's a power supply issue.
  2. Yep, that sounds like a race condition. Have you tried just setting the env variables in python (putenv)? When Houdini spawns MPlay, MPlay inherits all the environment variables that were set. Otherwise I suppose you could wait until the .flipbook_lock file appears, and then run the flipbook, as it's MPlay that creates the lockfile.
  3. You can determine which mplay sessions Houdini is linked to for flipbooking and rendering by looking at $HOME/houdiniX.Y/.flipbook_lock.<hostname> and .mplay_lock.<hostname> (if they exist). The first number is the process ID of the connected mplay session.
  4. Viewport Display Hair

    It's being looked at. In the meantime, turn off SubD curves on the hairgen objects.
  5. Highlight in viewport ?

    It's in the Optimization tab of the Display Options.
  6. Highlight in viewport ?

    The shadow is the individual bounding boxes of the instances being culled. It looks like a solid blob because they're all overlapping. This is a feature that prevents instancing from generating a massive number of polygons to draw, which can slow your redraws down (and if you're on Windows, cause the driver to reset). You can tweak the Scene Polygon Limit to balance redraw speed with the # of instances actually drawn.
  7. Persistent visualizers

    Create the visualizer in the Common section, then in the Display Options click "Save as Default". I would suggest doing this in a fresh session to avoid any display option contamination from a hip file.
  8. Vector maths basics

    You can also think of vectors in 1D - just a simple number line. 2+3 = 5. You start at zero and go to 2 (A). B then goes 3 units to the right (2->5). So the total vector is 5. You original example just extends that number line into 2 dimensions (2 number lines. which you can solve independently), and if you're dealing with 3D vectors, then you get an additional number line.
  9. Tensor cores are only good for doing 16b FP matrix multiplication, which is only useful if you're running some sort of machine learning algorithm that uses this sort of neural net processing. So it could be good for noise reduction. The RT cores appear to be only generally accessible via CUDA, DX12 or Vulkan (or via Nvidia's OPTIX library). As for the VRAM, anything bigger than 11GB appears to be reserved for the Quadros.
  10. New PC only for Houdini

    https://techreport.com/news/34008/nvidia-announces-geforce-rtx-2080-ti-rtx-2080-and-rtx-2070 Unless you're replacing a very old card, I'd wait to see what kind of support RTX has in the wild. Be a shame to have all those RT cores sitting idle. Edit: replaced with official link
  11. about flip_book

    If you mean into 1 movie, no.
  12. need advice on PC components upgrade for Houdini

    Have you noticed that the viewport is sluggish with the scenes you're using? If not, you can limp along with the 960 for a while until you notice it becoming a bottleneck. If you're going amd, the threadripper 2950X looks like a great value. If that's a bit rich for you, a Ryzen will do just fine. I'd recommend 32GB for a new system.
  13. Houdini 17 Wishlist

    Nvidia's Optix library is probably the most accessible way to get at those new RT cores from a non-GPU renderer. CUDA targets the existing shader cores, though I suspect there's probably a CUDA lib which handles dispatching to the RT cores (there is one such lib for dispatching to the Tensor cores).
  14. Ah, the days when a $5000 Quadro 6000 seemed like a lot...