Can you guys once day make fire that looks good but resolution independent like basic 3ds max fire? That way you can get as close to it as possible without have to resimulate fires/smoke close ups. Or fire generated during rendering with always sharp image depending how close your camera is from fire I think basic 3ds max has procedural fire but it looks bad.
Not sure how e-on Ozone 6 made procedural nice looking clouds during rendering.
By the way will we ever see where PHoenix can use multiple CPUs for calculation kind of like what we do with octane network rendering 20 GPUs per one frame at the same time?
the funny part is that the very very first phoenix (without fd) was able to do this, it was not a simulator, but an atmospheric procedural fire effect. but i think the procedural fire is always worse than the simulated one.
Hi yes spread sim over multiple CPUs. But I think GPU calculations would be great. We use Lucid by Ephere sometimes and it uses Nvidia Flex and it is fast except liquids dont look good at all.
On the resolution independent thing you won’t get much choice with modern simulators since they all use a grid / voxel method similar to pixels in your image. For something you can get infinitely close to you’d be asking current tech to do a hell of a lot of work for very little gain, you’d probably run out of ram too
What a lot of people do is use something like krakatoa and put millions of particle through the movement information of a fume or phoenix sim which has the effect of uprezzing it a lot, it’s another plugin though.
By the way will solid state hard drive improve phoenix sim calculations compared to HDD? I read somewhere it is not just CPU and memory that matters but also harddrive.
Also if solid state harddrive will make it faster can i have 3ds max and phoenix on my regular HDD and sim files will be recorded to a very fast solid state harddrive?
in case when you export a lot of particles (they produce bigger cache files), a ssd will certainly improve the performance, our experiments show that with 50M particles the bottleneck of the simulation is exactly the cache saving
Can you elaborate more on the multiple CPU question?