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 malexander

  1. It's sort of possible. You can write a HDK scene hook which completely takes over the beauty pass, then renders the scene 5x with 5 different projections to a cubemap. Then you'd take that cubemap as a texture input to a GLSL shader which does the fisheye lookup and writes it to the beauty pass framebuffer. I say "sort of" because the rendered image will be out of sync with everything else in the viewport, such as picking, handles, construction plane, etc. so in my mind it's not really a valid solution (more of a hack). This is probably something that'd be better in the OpenGL ROP or the Flipbook, which is less concerned about those other things.
  2. You could be running out of file descriptors ("Too many open files"). On Linux, you can run 'limit' (or ulimit) to see how many files can be open at once. On my system, it's 1024 by default. Which OS are you running?
  3. There's also Ce (emission color), Ca (ambient color), and Cs (specular color). Only Cd is supported by the viewport, though. For the principled shader, Cd is considered the base color (PBR rules).
  4. The menu item does work for me - What desktop are you using, or is it a custom desk?
  5. The easiest way to do this is to click on the arrows in the middle of the split bar:
  6. No, you pretty much can't do this at the moment. Large angle FOV projections will severely mangle geometry with large triangles, so it'd have to render to a cubemap, then sample from there into a regular 2D image using a lens shader. I don't think that you could accomplish this with the HDK right now. You could render a cubemap with the GL ROP in 16.0 and post-process that, though.
  7. If they ran the test with the CPU-CL driver, I'd expect the 1800X to slightly edge out the 1700X. But even then, it'd be "roughly the same" in that you as a user wouldn't notice the difference unless you were sitting there with a stopwatch
  8. When you're simming or rendering, you'll notice those extra two cores. If you do that a lot, 400eu will pay for itself in a short fraction of the CPU's lifetime.
  9. On a 32" 4K you could also try Large UI Size, if you find High DPI is too large.
  10. Yep, you could run a bunch of them in 16x mode, and have some lanes leftover to access a couple of PCI-Ex-based SSDs for large datasets. This is particularly important as AMD GPUs now have virtual addressing which could target the data on SSDs directly (though I'm unsure if that's currently supported for external SSDs, or just the TB one built in to the new Radeon Pro SSG cards). Usually there's a few lanes taken up by ethernet, the chipset, and slow storage as well, so 40 can go really quick.
  11. Looking at the consumer chips, they have a dual DDR4 interface which is faster than the 4-core Haswell and lower Intel GPUs (~43GB/s vs. 25GB/s) but slower that the newer SkyLake+ CPUs (50+GB/s). The quad-channel socket 2011 and Xeon chips leave them in the dust at 75+GB/s. That could be a potential bottleneck for very large sims which require a lot of mem bandwidth. I think this is probably the weak link in the Ryzen design. A 16 thread CPU requires a lot of memory bandwidth, and it could be starved by a dual channel interface. The server chip doesn't have this limitation, but it also takes a clockspeed hit.
  12. While it's great that AMD is competing again, going for a Zen-based chip is more of a cost decision than a performance one. Hopefully it'll put pressure on Intel to drop prices a bit over the long haul.
  13. Should be fixed now in 16.0.555. Rookie mistake :-/
  14. In H16 we finally dropped the requirement that C and A must exist regardless of the actual file contents. This is good in a lot of cases (why create C and A when an EXR contains only Pz) and was pushed by the Terrain project. It was also the original intent of COPs, so I'm glad that this change finally happened. Unfortunately, due to the fact that this restriction has been in place for a dozen versions or so, a few cases that expect A to exist have now broken, the lumakey and chromakey among them.
  15. It wasn't loading the reflection map size properly from the display options save file. As of build 552, set it to 512 and Save as Defaults again, and it should be good to go.
  16. Yes, you need a light in order for HQ lighting to engage. 16x16 seems like an awfully low setting though, not sure why it'd be set to that.
  17. Reflections requires HQ Lighting, Reflections enabled, and a material with the GL reflect parm/tag set to a value greater than the minimum reflect display option. As of build 550, Metallic will be used if the GL reflect parm isn't found, so Principled shaders will work out of the box for metallic objects. It works by removing the reflective object and rendering a cubemap of the scene at the objects origin using normal quality rendering. As such it can't do self reflection or multibounce reflection between two reflective objects. There is also some minor distortion at the edges of the reflective object for geometry that's far away from the origin. It seemed pretty stable when I tried it, so if you have a case that's not working please submit a bug.
  18. The Classic Shader gives you more artistic control over the various lighting components. The Principled shader sticks to PBR principles, and thus there's no parameter to adjust specular intensity.
  19. We do have a lot of scenes from the demo material produced for the releases which could double as benchmarks, though they'd take some finessing and cleanup. From what I've seen, it would also be rather large in terms of a download. Some of the fluid sims and hair/volume renders would be good candidates. Also, we're not really in the business of promoting hardware
  20. To my knowledge, we don't have a Ryzen system to test with yet at SideFX. And while you can buy a Ryzen CPU, finding a motherboard to plug it into is difficult - Ryzen motherboards are apparently sold out of most online stores Ryzen is fairly competitive compute-wise, but the one thing that might hold it back compared to a 6-8 core Intel CPU is memory bandwidth. Ryzen has a dual-channel memory interface (42-44GB/s) vs. Intel's quad-channel interfaces (68-74GB/s). You might see a hit in large sims that consume a lot of memory. That'd be one of the first tests I'd try.
  21. You can also try installing the Nvidia 378 driver series, if you're using an older driver. http://www.nvidia.com/Download/Find.aspx?lang=en-us
  22. You should download the Qt5 version unless you have some particular need for Qt4 (such as a lot of Qt4 add-on scripts, which doesn't appear to be the case). For bug fixes, you'll want the Daily Build page found here: https://www.sidefx.com/download/daily-builds
  23. Download any H16 daily build and this should take care of the viewport sluggishness. Others ran into issues with the UI being sluggish and upgrading to the nvidia 378 drivers (from 376) really helped.
  24. "Hyperthreading slows workloads" is from the days of the old Front-Side Bus, which couldn't supply the CPU with data fast enough to cope with double the threads. That changed a long time ago when Intel ditched the FSB with the Nehalem architecture, when the i7/5/3 product prefixes came into being. Now you generally only get a slowdown if threaded code scales poorly, if lock contention begins to dominate over real work. Running houdini -j<n> to restrict Houdini to use fewer threads in those cases (such as the cases Jason mentioned) would be a better solution that disabling it unilaterally.
  25. Have you tried a daily build of Houdini 16? Several viewport performances issues were fixed in the days leading up to the release which didn't make the gold release build itself.