German website PC Games Hardware has posted some videos showcasing an ILM presentation on their GPU fluid simulation tool, PLUME. The website is in German, but the videos are in English, and you can check them out over at PC Games Hardware.
CGPress is an independent news website built by and for CG artists. With more than 15 years in the business, we are one of the longest-running CG news organizations in the world. Our news reporting has gathered a reputation for credibility, independent coverage and focus on quality journalism. Our feature articles are known for their in-depth analyses and impact on the CG scene. “5 out of 5 artists recommend it.”
Man how much faster those tasks would be to iterate. I feel the 3rd party VFX world is a little bit behind on exploring the potentials of GPU. I understand it’s difficult programming, but even just getting PhysX’s versions of gaseous/solid liquids and hair in 3ds max would be awesome.
I use GPU Vray all the time to test out lighting and it’s so fast compared to CPU even if I can’t fit all that much texture on my current card. Even upgrading to a 3GB card wouldn’t be that expensive and that would allow you at least a bit more. Then for companies, getting a 12gb card isn’t out of range considering how much time it would save. Plus it frees up the CPU so you can continue doing some tasks that aren’t display intensive.
I am grateful that our viewports are a lot faster these days because of GPU technology but it really would be awesome if it was used for more things like this.
How difficult/expensive would it be to implement these new PhysX features into Max?
In the times when it was becoming clear, that GPUs can do various parallel computations I’ve heard different opinions on the same matter from some really top programmers from the industry. One camp was dismissing the whole idea saying mostly that some tasks are impossible (turned out they were possible with just the next generation of GPUs and SDKs) while the other ones were very very enthusiastic and saw the challenges of coding for a GPU as more of a fun challenge than some torture as the first camp saw it.
So it is what it is – programmers even though very very good are not able to predict the future of Hardware and those dismissed the GPUs are paying the price. Those who took the risk have a serious advantage (e.g Otoy).
So ladies and gents.. when it comes to the future of the HW, never turn to a programmer no matter how good he is. Turn to him when enoucntering coding issues but never when you’re encountering strategy issues especially when their tied to a future gen tech.
Agreed, but compute on GPU isn’t exactly that new anymore. I’m getting a first hands look tomorrow at the PhysX stuff and I’m sure it’s not as developed (at least not for VFX) as ILM’s but I still feel it could be very useful for iterating. Hoping to be able to ask them if I can spread the word a bit, maybe make a video.
Maybe the Chaos Group guys will get Phoenix to get all the computing on the GPU since they are already in that field with Vray. I know FumeFX was getting it but not sure what the future lies with that and how fast it will be compared to PhysX and the like.
Let us know how the first hand of the Physx demo goes. From what I heard it scales up really well and can be used in vfx not just games.
Impressive to say the least.
The fourth video has nice fly-through examples, man they look good.
Shocking how quick they are producing iterations, he said what I thought was 13000 sim passes in 7-8 months on transformers! WTF