how many particles is too many?

[thebigmontagestudios]

I'm working on a sim using a particle based scene. I didn't expect it to be as large as it evolved into but at one point the sim reached about 8 million particles. When i meshed it out, it was around 4-5 million faces. It was supposed to be a 144 frame sequence and around 65-70, after hours of simming, twice my system has frozen by saying that i've run out of application memory. I have heard that this warning is a Mavericks common issue these days. I am running it on a latest iMac with a 3.4ghz quad-core and 32GB ram. It should be a pretty powerful system but this scene keeps killing it.

So two questions i guess. Should i NOT be using particles for a sim this size and instead be using Hybrido? Typically, is there a rough number of particles that 'sounds like' too many to use before crashing a scene? I'm sure Maya would probably struggle running the sim through at nearly 5 million faces on the mesh too. Any thoughts on this?

Out of curiosity, as I will be upgrading to the beast of the Mac Pro in the near future, would that bump in cores and processing power make that big of a difference in a situation like this or should my workflow be different no matter what?

[gus]

Hi Mark,

I'd say that the main issue you are facing is the fact that you have a big amount of sph particles, and the .bin format is a heavy one. We will revise that for future versions of RealFlow, but I'd say that you would be better running big sims with Hybrido if possible. The rpc format is optimized and likely to replace all .bin.

If you have to be running with SPH, for whatever reason, try to split between multiple emitters, again if possible, so you have less particles at a time in memory.

Your machine is ok with 32GB, and changing to a newer one with the same amount of memory won´t benefit the current issue you are having to be honest.

Cheers,

g-)

[thebigmontagestudios]

yeah it's been a while actually since i really did anything with the SPH solver. And i wasn't originally thinking it would be as many particles as it turned out to be. Gonna give it a go with the hybrido and see how that works out.

Usually, is there like a rule of thumb on what to cap out the number of sph particles in a scene at? Obviously, it will differ based on hardware and stuff, but in general, is there a number you might have in mind?

Is that the plan that everything is eventually going to run through the Hybrido system and the particle based is going to go away? Just curious

[tsn]

Just as number: Fusion CIS made a SPH shot for G.I. Joe in 2010 and the scene had more than 25 million particles spread over 9 emitters.

http://www.fusioncis.com/press/gijoe/su ... hotcam.mov

Maybe you'll find more information about this shot was made here:

http://www.facebook.com/home.php?ref=ho ... 6394379990