I have this quite simple Phoenix FD scene, but I can’t render it. Max crashes to desktop without any explanation. I guess 128GB RAM is not enough for a simple simulation to be rendered. What are Your recommendations for reducing RAM usage during rendering? What is the minimum RAM required by Chaos Phoenix.
Please, send the scene so I can troubleshoot.
Thanks,
I’m using built-in Speedboat setup. Everything renders fine up to frame 67 and after that nothing renders. I’m rendering to resolution 640x480.
Are there any instructions on how to tweak speedboat settings or how to build a new from scratch?
Any errors in the log? Have you monitored your memory usage while rendering? Otherwise, here’s a guide about active bodies. Otherwise, here are the hardware recommendations for Phoenix and V-Ray. The required memory is, of course, scene dependent.
Thanks,
Here’s what happens if I try to render the scene. Actually it’s not using that much RAM before the scene chrashes.I don’t use Activebodies with boats. I only take the simulator setup and animate boat manually. I new got the boat simulation correctly with Active bodies. The boat jumps a few times and the dives.
I think it’s the foam, that causes crashes. If I reduce the foam multiplier to 0.5 I can render up to Frame 131 (previously 67).
It doesn’t look like a memory issue. Best you send the scene.
This is very generic scene I’m using search for decent settings for boat simulations. It’s based on speedboat setup. The boat I’m using comes from Turbosquid so the license would not allow sending the scene.
I think, the source for crashes comes from Phoenix foam particle shader. Just need to find a way to get decent amount of particles without crashing. At the moment I’m playing with the half life settings to keep the foam particle amount renderable.
E: At the moment I’m working on water jet settings. I have created liquid source and a cylindrical emitter for it. However nothing is created. Can anybody tell how to create that. In generally I would like to see foam coming out of waterjet nozzle with high speed, but I don’t neither water or foam.
Hello,
You mentioned that you are using a model purchased on Turbosquid. Generally, it is a good workflow practice to isolate the hull of the boat, make a copy, then optimize the geometry and, more importantly, fix any mesh errors, such as open edges, duplicated faces, and so on. If you had an optimized collision proxy, you could’ve sent the scene without the Turbosquid model. Use the proxy mesh to drive the simulation, and use the original turbosquid model for rendering only. Phoenix attempts to voxelize every piece of geometry it interacts with. If the mesh has any of the errors mentioned above, it may fail. It may not happen right away, but potentially it might cause issues. Especially if the model is animated and rotating. Check the cache content window for any rapid changes in velocity, for example. Try rendering the sequence (or the problematic frames) without motion blur, displacement, or with the particle shaders turned off, to see if you can isolate the issue.
Regarding the water jet scene, you need a source with the “Emit liquid” checkbox turned off and set to emit the desired type of secondary particles (foam, splash, mist). You also need to have the type of particles you want to emit enabled. Can you show a screenshot of your setup?
Here’s a demo:
Cheers
That’s great, but I’m not sure how I can help. Can you send a scene containing only the Phoenix simulator, the source node, and the emitter cylinder to https://support.chaos.com/hc/en-us/requests/new ? I don’t need the lights, textures, or any other geometry.
I got the scene, simulated, and rendered 81 frames without crashes or any other issues. Screen capture - c9911caf82e6603b3e4a87f6680260cf - Gyazo
I noticed, though, that the model has 1254 open edges. Screenshot - df34b26da96a78aee52ebaf0cf9ca78a - Gyazo
This might not be an issue in this case, but, as a general rule of thumb, all objects that interact with Phoenix should be closed, watertight models. Phoenix does a pretty good job solving messy geometry, but you never know when the algorithm could fail.
Avoid objects that are thinner than a single voxel. Add a push modifier or delete them entirely if they do not contribute to the final look. ( Screenshot - 527a8666f021de0cd228931e4459f84d - Gyazo )
I don’t think you need these (Screenshot - 21ee24e9bdc9b495541a0f09c3861151 - Gyazo), or these ( Screenshot - 153d379451b9dab0a8b6050e5d59bbb5 - Gyazo ), and definitely not these ones (Screenshot - 1a2bba8147fbf00f4c9cc6f50447c336 - Gyazo). Tiny screws, smaller than a voxel, won’t contribute to the simulation, yet Phoenix will still go through all the triangles trying to resolve them.
But anyway, this was rendered on 64GB RAM, so the lack of memory isn’t the issue here. Which versions of Phoenix and Vray are you using?
Thanks,
I’m using 5.24.01 Phoenix and VRay is 6 update 2.1,
Was there any reason why the Waterjet emitter is not working? I’ve done waterjet simulations number of times, but this is the first time I see this issue.
Also I think, the open edges are not affecting this simulation, because the hole in the ocean matches boat hull.
I also noticed, Initial Fill up and Ocean level have slightly different values. This will explain few things.
If you mean the cylinder that emits foam particles, it seems to be working here. Is there something specific that you want it to do?
Yes, I noticed the difference in level between the ocean fill and the ocean surface. It is visible in the render. Was that created by the preset?
This is an expected behaviour. Most of the secondary particles (foam, splash, mist) are allowed to exit the grid and live outside of the simulation for a short period of time. However, once they leave the simulation area (grid), they behave as “dumb” particles, and unfortunately, there’s not much you can do to control them. Most likely, the foam outside the grid is not displaced properly, since there are no liquid particles around it.
Try disabling the displacement temporarily to see if this is the case. Or you can hide the grid and render only the foam shader.
To avoid this, you have to:
1. Make the grid longer.
2. Make the foam disappear faster.
3. Try to increase the vertical offset in the particle shader to see if the foam will pop above the water. This will raise all the foam, including the bubbles inside the grid. However, if this fixes the appearance of the foam outside, you can use a cutter object and a second particle shader. Essentially, one normal shader for the inside particles, which is cut where the bubbles leave the grid. And a second one for the foam outside, which has the vertical offset.
Personally, I would try one of the first two options.
Were you able to render your sequence? Both Zdravko and I rendered all simulated frames without issues.
I tried to generate a new version of the simulation. However Phoenix had crashed at frame 356. Is there anything. Then if I start a new session, go to frame 356 it works fine. Then I move backwards to Frame 355 Max crashes. How do I debug this? The AUR file size at that point is 181 MB.
Attach the dump/log files so the devs can have a look.
Any ideas for the black spots on foam?
I created yet another simulation, but the spots are there. They seem to be one voxel sized waterdrops that render black within foam. Any suggestions?
Use the link I shared. Here it is again: How to Report a Bug
I’m not an expert in Phoenix, which is why I cannot help you with the particulars, and why I asked Zdravko to help you (he’s part of Chaos, too). I’ll ping him to take another look at the thread. But bear in mind - Phoenix’s status was recently changed to end-of-support.