![]() |
|
Fume fx scene loads-explosive initial sims anomalies?! - Printable Version +- Sitni Sati Forums (https://forum.afterworks.com) +-- Forum: Products (https://forum.afterworks.com/forumdisplay.php?fid=4) +--- Forum: FumeFX [maya] (https://forum.afterworks.com/forumdisplay.php?fid=19) +--- Thread: Fume fx scene loads-explosive initial sims anomalies?! (/showthread.php?tid=1553) |
Fume fx scene loads-explosive initial sims anomalies?! - rconover - 07-29-2014 This is something I've only seen in Fumefx maya (linux). Never on MAX. But I've noticed it in many scenes now-so it seems like something is not quite 'cleaned up' right when you load an existing fumefx in a scene. Possibly it has to do with sim start frames that are not 1 or 0-but who knows?! I'm not certain if its always the initial scene loaded in the maya session either-but it might be. What you get- You load an existing Fumefx scene-begin a sim- and the first results look exploded-like your velocities of time scale are set to a rediculous amount. Also it can sometimes drive the memory up exponentially as Fume FX 'tries' to calculate some crazy exploded values, grid sizes as adaptive grid tries to keep up. Maya may crash too! I've experienced this with setups that i am familiar with-and know these 'anomaly sims' make no sense. I simply stop the sim. And start it over again. Now the result is as expected. The 'anomaly sim' won''t return unless i exit maya and load the scene again. I've checked with colleagues (on maya linux) and its true for them as well. Not so bad as long as you work on your own machine and catch it happening. BUT I could see this being very problematic doing sims on a farm-forcing users to sim on their own machines. This could be a show stopper if they need to farm sim. Repro- -This will never happen with a brand new Fumefx added to a scene. It must reside in the loaded scene already. -It always sorts itself out correctly the *second* attempt at doing a sim. -Possibly related to having Maximum Simulation Steps to other than '1'. I often set to '3'. Can you repro with any of the scenes I've sent you in the past? If not I'll try and make you one. I'd also like to know if anyone else on maya has experience this. Possibly it could be our pipeline and the scene presets we load maya with. Re: Fume fx scene loads-explosive initial sims anomalies?! - dagoose - 07-30-2014 Hi rconover, I discovered something similar with open boundaries and the container not sitting in 0,0,0. The sim explodes in the first frame and fumeFX fails to allocate the neccesary memory ... which is probably super hugh due to the open bounds and thats it ... maya freezes. When you close your bounds .. sim the first frame .. open it up again ... then the sim runs as expected. Greets Rico Re: Fume fx scene loads-explosive initial sims anomalies?! - sitnisati - 07-31-2014 Hello Rico, That is the issue as described here: viewtopic.php?f=16&t=1616 It is fixed for 3.5.5. Thank you! Kresimir Re: Fume fx scene loads-explosive initial sims anomalies?! - sitnisati - 07-31-2014 Hi rconover, We are looking at this issue... thank you for a bug report. It should be fixed in 3.5.5, but still we don't know the cause of such behavior. Thank you. Regards, Kresimir Tkalcec Re: Fume fx scene loads-explosive initial sims anomalies?! - rconover - 07-31-2014 dagoose Wrote:Hi rconover,Indeed those are factors for us too (and most production scenes-I would think). I am glad it looks like its fixed. Yeah! Re: Fume fx scene loads-explosive initial sims anomalies?! - dagoose - 08-05-2014 I further realized that if the container is not close to 0, 0, 0 also the continue sim doesn't work properly. It sims a couple of frames and then it residing in calc and does not come back. After moving the whole sim close to 0, 0, 0 the continue sim works. cheers Rico Re: Fume fx scene loads-explosive initial sims anomalies?! - sitnisati - 08-06-2014 dagoose Wrote:I further realized that if the container is not close to 0, 0, 0 also the continue sim doesn't work properly. It sims a couple of frames and then it residing in calc and does not come back. Hello, That's been fixed in 3.5.5 as well. Thank you. Kresimir Tkalcec |