Welcome to od|forum

Register now to gain access to all of our features. Once registered and logged in, you will be able to contribute to this site by submitting your own content or replying to existing content. You'll be able to customize your profile, receive reputation points as a reward for submitting content, while also communicating with other members via your own private inbox, plus much more! This message will be removed once you have signed in.


  • Content count

  • Joined

  • Last visited

  • Days Won


Everything posted by lukeiamyourfather

  1. I would go with a ThinkPad P50 if you absolutely must have a laptop (up to 64GB of memory and Xeon processor option). I wouldn't go with a Mac for Houdini under any circumstances. Apple and their "pro" nomenclature is a farce, 16GB is simply not enough and you can't upgrade it. Don't bother with a laptop unless you have no choice in the matter because you're throwing money out the window for mediocre performance and limited scalability.
  2. Exactly what Mark said. The extra PCI Express lanes are also going to be useful for high performance storage clusters with NVMe arrays or machines with tons of drives like ZFS. All around it's going to be a very useful platform in CGI production if the pricing is competitive.
  3. Yes, the GTX 1060 is supported in Linux. If possible pick a distribution with a repository that hosts the Nvidia drivers and associated DKMS configurations. One example is Ubuntu, they do a good job with the proprietary Nvidia drivers. https://help.ubuntu.com/community/BinaryDriverHowto/Nvidia A common stumbling block for users new to Linux is to download drivers from a manufacturer and try to install them. This is what you do in Windows. This is not what you do in Linux (unless you hate yourself and want to have a miserable day when the kernel updates and kills your drivers). Only install software provided by the repositories unless there's a good reason like the software isn't available in a repository like Houdini.
  4. Naples is targeting supercomputing and massively parallel applications (Mantra and other renderers are definitely in this category). It will be especially useful for OpenCL and CUDA with 128 PCI Express lanes, that's crazy. The Xeon E5-2600 series has 40 PCI Express lanes. It may or may not make a good workstation platform depending on how much stuff you do that relies on a single processor core. These days that's less and less common but there are some applications that lag in various areas. It will likely make a good render node and simulation node if their pricing is competitive like the Ryzen products.
  5. The dual Xeon machine will be faster and it can accommodate a lot more memory. The downside is it'll require more than twice as much power, this can be quite annoying in a small office or home setting because it'll make the room hotter.
  6. Major upgrades have compatibility breaking changes (e.g. H15.5 to H16.0). This has always been the case. Old shelf scripts, digital assets, scenes, and the like are not going to work 100% of the time in another major version. You'll have to asses the things one at a time and make necessary changes if you want to bring over all of your previous stuff. This is why production facilities will use the same major version of Houdini (or any other software for that matter) for the length of a production even if it's outdated by the end of the production.
  7. That would be nice but I'd rather the development time be spent on features we can use in production. That's my two cents.
  8. On Cinebench R15 the 1700X is on par with a 6900K ($399 and $1089). The memory bandwidth might be an issue but I see the number of memory slots a bigger issue (4 versus 8). It's as if you could get a cheap 8 core processor for the 1151 socket platform in terms of what it's offering. For students and hobbyist this is awesome. For professionals I don't see it as that alluring. I'm hoping their Opteron refresh based on Zen will bring some competition back to the server and workstation market.
  9. There's a viewport performance issue with the first public release of Houdini 16. It has since been fixed in the daily builds (as discussed above). Below is the journal entry with at least one of the issues being fixed. https://sidefx.com/changelog/?journal=16.0&categories=&body=viewport&version=16.0.528&build_0=&build_1=&show_versions=on&show_compatibility=on&items_per_page=
  10. The latest daily builds have addressed the viewport performance issues. For example the link below for the Windows daily build (16.0.541). https://www.sidefx.com/download/download-houdini/48682/ If you're not familiar with daily/nightly build concept please read this link below. They're generally not as reliable as the stable releases but they might fix bugs you're encountering and get you by until the next stable release. https://en.wikipedia.org/wiki/Daily_build
  11. You don't need to download anything extra to install Houdini. Qt is a framework for building graphical interfaces for software that only developers need to have installed. If you don't know what this means then just download Houdini 16 installer at the top of the download page (or one of the daily builds for Houdini 16). The Qt4 version at the bottom of the page is for studios with tools built with Qt4 integrated into Houdini. Again, if you don't know what this means then just ignore the Qt4 version completely because you don't need it.
  12. Someone probably left the team or someone new took over work from someone else. It's amazing how much hinges on just a few people in software and technology companies. Windows doesn't inherently support OpenCL so you have to install software to use it. AMD provides an OpenCL runtime for their processors in one of their SDK downloads. This is might be a moot point if you have a GPU that also supports OpenCL. http://developer.amd.com/resources/articles-whitepapers/opencl-and-the-amd-app-sdk-v2-4/ http://developer.amd.com/tools-and-sdks/opencl-zone/amd-accelerated-parallel-processing-app-sdk/
  13. The way that makes the director happy!
  14. By default most Linux distributions use the Alt key plus clicking as a shortcut for dragging a window without having to click the title bar of the window. http://askubuntu.com/questions/118151/how-do-i-disable-window-move-with-alt-left-mouse-button-in-gnome-shell I use GNOME and that's how to fix it in GNOME.
  15. I've used Houdini on Linux for almost a decade. It wouldn't surprise me if most Houdini customers were using it on Linux. Most of the time I've been using Ubuntu and sometimes Fedora and CentOS. Are you using a high DPI display like UHD or a Mac with a Retina screen? Have you used Houdini on other platforms or is this your first time using Houdini? Are you using the graphics drivers from the manufacturer or the open source drivers?
  16. There's some information about it in the Houdini documentation. http://sidefx.com/docs/houdini/render/understanding You change the mode on the Mantra node. http://sidefx.com/docs/houdini/nodes/out/ifd#rendering The setting that determines how much to subdivide things is called dicing. It can be global, per object, based on distance, all kinds of fun stuff.
  17. If you're using micropolygon rendering you can displace and subdivide things practically infinitely. Generate the displacement map at render time so there's not a resolution limitation involved. I haven't tried with other render modes but it might work in other render modes as well.
  18. In the past I've used Time Shift > ROP then back in with File > Time Shift. Multiply and divide things by whatever with the Time Shift and it doesn't matter what the frame numbers are on the ROP and files. If you're doing DOP stuff it saves out the substep by default.
  19. Open a terminal, set the variable, then launch the Houdini instance. Exactly how you do this depends on the platform you're using.
  20. Unless things have changed recently you can use only one device. Though you can load other instances of Houdini with different OpenCL devices and run them concurrently working on different simulations. See the documentation for the environment variable to define the device to use.
  21. That's a good resource, those were the kludges I was thinking of (using kludge because we're talking about rendering something rather than making a video game). You still have to render a bunch of stuff to make that work though which is what the original poster is trying to avoid. If you're going to take the time there to render the sprites you might as well just render it in Mantra outright. The payoff isn't really worth it for one time use versus a game where it would get many uses. The vertex animations in Unreal Engine in rely on consistent topology. Something like a FLIP simulation would have arbitrary meshes on each frame. To animate arbitrary meshes you'd need a different solution like toggling their visibility over time which has it's own set of issues. I get the desire to render things quickly but Unreal Engine isn't a renderer it's a game engine so the workflow and optimizations are all geared towards that. I'm using it as a makeshift renderer on some projects but it comes with a lot of strings attached.
  22. You're better off rendering fluids and simulations in Houdini with Mantra. I use Unreal Engine and Houdini on a daily basis and there's just no clean way to render volumes and geometry sequences. There are some kludges out there but why bother if you're starting in Houdini to begin with?
  23. Needs more cowbell particles. If you're at the limits of your machine consider narrowing the scope of the simulation. For example only simulate what is directly around the character versus everything everywhere. Then combine meshes or simulations later in the pipeline so they seem like one.
  24. It looks like operating system level interface scaling for a high DPI display. The scaling doesn't work good on applications that are rendered with hardware (OpenGL) like Houdini. Houdini has it's own way of managing high DPI displays, click Edit menu, Preferences, General User Interface, and then change the Global UI Size parameter.
  25. You could use something like sed to add stuff after the fact though I think you're better off using a wrapper script that sets things up before calling the render.