Tweet
The crash is back. This time I got to see the crash error that turns up, and its actually the same error about the globalillumination thing that we have this hotfix on. Delete the atmospherics, save and restart. But when I open the scene to delete the atmospherics, its not there.
-
-
the tread was started with report for crash when foam/splash are generated, this does not sound like the same situation, rather as a rendering related crash. anyway, we have to fix any crashes so give us the log file to see what is happening______________________________________________
VRScans developerComment
-
-
I tried to start a fresh scene. Import everything one by one.
I made a new phoenix fd grid, and imported the cache files. It rendered.
I made a material with the foam tex, and disp tex, it crashed.
The material is a blend material just like the one you use on your example scene on vimeo when showcasing the foam texture.
EDIT: The old fix for that crash does not work on this scene. I have tried to both import everything to a new scene, and fix the old scene. Deleting the atmospherics (is no longer there) and the phoenix part in the enviroment rollout. It crashes right as I try to link to the cache files I have used 30 hours to calculate and hit render.
EDIT2: Here is one of the cache files if that helps.
http://www.filedropper.com/nd83068ph...dliquid0010732Last edited by oophus; 16-10-2015, 08:50 AM.Comment
-
yes, in the log you attached is visible that the crash is caused by the frame loading (or maybe preview), it happens immediately after the frame load.
we will investigate it, hopefully the cache file will help______________________________________________
VRScans developerComment
-
-
you can try, in the worst case the backup interval will keep your result. btw, when you have a crash, it's not a bad idea to try restore from the nearest backup, if the crash happens again, we can fix it starting from the same backup frame______________________________________________
VRScans developerComment
-
Because of the last render bug which caused flickering in rendertime because of the backup interval, I did not use that function during this simulation.Originally posted by Ivaylo Katev View Post
Is it safe to calculate with a backup interval now, as in the flicker during rendertime is fixed?
EDIT:
I have installed the new nighly from today, and it renders like a charm! Thanks a bunch for this insane, quick and superb support!
Last edited by oophus; 19-10-2015, 02:03 AM.Comment
-
i completely forgot about the flickering, that is really important. i still think it's a question setup, but have to find what is causing it and to display a warning______________________________________________
VRScans developerComment
-
Ah ok.Originally posted by Ivaylo Katev View Post
On another note. The scene renders now, but when trying to resim it, it crashes.
I made a complete new scene just to test. Also to test foam and splashes to see if they stick to the container walls, but it crashes during the simulation. I made a complete sim through with 15m grid. But at lower grid size, it crashes.
Here is the test file:
EDIT1: And now after reopen the scene and trying again, it simulates through...
But it looks like the particles doesnt leave the container, which is a problem. (some of them do, but the majority does not).
EDIT2: Stopped the simulation. Took down the grid size to about 2m, and also upped the steps per frame to 4 from 1, and it crashed at first frame.
EDIT3: Upped the resolution to 17m, and it calculated through. Took it down again, and it crashed. Took it back up to 17m to doublecheck and it crashed.
I cant find a clear case where this crashes or not. I also tried to disable all of the preview functions during sim.
EDIT4: I get a weird result when pressing "show mesh" in the preview window.
As shown, the ocean mode makes mesh within the entire grid. The initial fillup is set at 20, the same as ocean level.
When trying to render this, just to see if the render returns a simular result as the viewport, the render returns a black screen. Now I'm really confused.
EDIT5:
Cap rendering mode also produces a weird result. But at least it renders, even though something is really broken at this point.
Last edited by oophus; 19-10-2015, 05:09 AM.Comment
-
ok, will test your scene, btw i tried to reproduce the foam-wall problem but without success, is this scene reproducing it?
now i think it may be an UI issue, because i see your simulator is the new one, i tested it with existing scene with the old simulator______________________________________________
VRScans developerComment
-
Yes, as far as I could see this scene did reproduce that. Most of the particles was sticking to the container while maybe 10% did go through and continued.Originally posted by Ivaylo Katev View Post
The scene is simulated with todays nighly, which fixed the render problem with the old cache, but when trying to resim it since I had the particle error, it crashes again.
EDIT1: Did a new take, and turned off the preview except for the particles. This time it calculated through, but the image may illustrate what I mean. At the top of the grid it goes through perfectly, and keeps its momentum. But on the side they stop, where most of the particles stick to the grids side, while others go through, but they slow down considerably.
EDIT2: Still the same weird problem with Ocean mode though.
Last edited by oophus; 19-10-2015, 05:45 AM.Comment
-
can you show the viewport with mesh switched off, to see the grid content?______________________________________________
VRScans developerComment
-
Not entirely sure what you mean, but putting on all the previews except show mesh?Originally posted by Ivaylo Katev View Post
Comment
-
That's the same issue I have with any of the recent nightly builds when trying to render on ocean mode. Ivaylo said it maybe something to do with frame load and was to be investigatedAdam TrowersComment
Comment