Jump to content
Solitude

More ICE Awesomeness

Recommended Posts

Pretty impressive....

edit: I still have to wonder what kind of sim times they're getting...

Edited by Solitude

Share this post


Link to post
Share on other sites

ahh you beat me to it,

i would say awesomeness is an understatement, this stuff is f*n amazing

i also wonder about the sim times, from what i can see on some of the stuff I'm guessing it's probably pretty decent

i'll try find out

jason

Share this post


Link to post
Share on other sites

It's so awesome ! amazing ! can that be re-written for Houdini ? :P

Pretty impressive....

edit: I still have to wonder what kind of sim times they're getting...

Share this post


Link to post
Share on other sites

Whoa, that´s pretty cool. The chunks of ground simulation at the beginning really blew my mind, but everything was quite impressive.

Share this post


Link to post
Share on other sites

Another, after Maya's nParticles, fully particle based physics engine. Massive integrator work but easy rules. I wonder if SESI has ever seen this as an option to dops.

Share this post


Link to post
Share on other sites

It's very amazing of course but my Houdinibrain is wondering whether we couldn't replicate some of these things with velfield advected particles and stickyness attributes based on a voronoi noise or something.

Edited by Macha

Share this post


Link to post
Share on other sites

It's very amazing of course but my Houdinibrain is wondering whether we couldn't replicate some of these things with velfield advected particles and stickyness attributes based on a voronoi noise or something.

But before doing the advection you need a good solver, and this is what Lagoa is doing really well.

the thing that really amaze me is that applications that compete with Houdini got most of their good stuff from 3rd party developers.

Many of the best stuff in ICE is done by other people not autodesk, same for 3DStudio Max.

BTW this video is trully impressive, and let you think if Houdini can compete with this.

Share this post


Link to post
Share on other sites

It's so awesome ! amazing ! can that be re-written for Houdini ? :P

I'd say, not right now.

Not because it's not capable, but it just couldn't compete with the speed. ICE was written from the ground up to be fast, like 8-12 months of dev of just optimized code.

But yeah, I was quite impressed when I saw how far Thiago had taken this, this was originally the SPH solver. Then taken to a way different level.

Share this post


Link to post
Share on other sites

I'd say, not right now.

Not because it's not capable, but it just couldn't compete with the speed. ICE was written from the ground up to be fast, like 8-12 months of dev of just optimized code.

Why not?

This is a solver written for ICE, not ICE.

You can write you own fluid solver, based on SPH, complitelly multithreading and using GPU in Houdini. It could be as fast as the stuff from Thiago.

My only concern is why Houdini still don't have a fluid solver much faster that the current ones.

Share this post


Link to post
Share on other sites

I also find moderately interesting that a big portion of Softimage´s user base (or maybe just the forums user base) are complaining about ICE being too technical and not artist-friendly...which was something to expect if you think about it.

Imagine not being used to Houdini and starting out with VOPs...not the simplest way, methinks :)

As a consequence it looks like ICE is starting to be regarded as a tool for the "big-boys" as opposed to the commmon user.

Edited by Netvudu

Share this post


Link to post
Share on other sites

Why not?

This is a solver written for ICE, not ICE.

You can write you own fluid solver, based on SPH, complitelly multithreading and using GPU in Houdini. It could be as fast as the stuff from Thiago.

My only concern is why Houdini still don't have a fluid solver much faster that the current ones.

I would like to think so too. But, also might there be some bottlenecks going into or out of the solver node if the entire simulation architecture is not multithreaded etc?

Also, his original version of SPH used only factory nodes AFAIK. Who knows how much of this relies on custom stuff?

I think ICE has some long way to go to match the scope of Houdini, but it sure is easy to use for solving certain tasks.

Share this post


Link to post
Share on other sites

As a consequence it looks like ICE is starting to be regarded as a tool for the "big-boys" as opposed to the commmon user.

And also has piqued interest in users (like me) for Houdini...

Share this post


Link to post
Share on other sites

I also find moderately interesting that a big portion of Softimage´s user base (or maybe just the forums user base) are complaining about ICE being too technical and not artist-friendly...which was something to expect if you think about it.

Imagine not being used to Houdini and starting out with VOPs...not the simplest way, methinks :)

As a consequence it looks like ICE is starting to be regarded as a tool for the "big-boys" as opposed to the commmon user.

But not every "artist" develops tools. Just not how things work in production. What is more common is someone smart makes a great tool, like Lagoa, and with minimal training, someone can roll out shots. ICE lets you do this and adapt it to your shot. Wasn't this easy ages ago.

In fact, it's not all that unusual in houdini-land. Someone sets up the networks, and someone else takes over with little understanding and is able to roll out shots. Not every warm body has to be able to write SPH solvers.

Seem like a bit of jealousy and trying to knock dicks in the dirt. Brilliant dev work is brilliant dev work.

-Lu

Share this post


Link to post
Share on other sites

I would like to think so too. But, also might there be some bottlenecks going into or out of the solver node if the entire simulation architecture is not multithreaded etc?

Also, his original version of SPH used only factory nodes AFAIK. Who knows how much of this relies on custom stuff?

I think ICE has some long way to go to match the scope of Houdini, but it sure is easy to use for solving certain tasks.

If you implement it in DOPs definitely you will need to deal with some DOPS bottlenecks. I dont know what these bottlenecks could be.

But my idea is to develop entirely in SOPs. You just read geo and solve it, just it. It is how many studios are developing their own solvers. And here you dont have bottlenecks.

Share this post


Link to post
Share on other sites

If you implement it in DOPs definitely you will need to deal with some DOPS bottlenecks.

I suspect the biggest bottleneck of DOPs are POPs. Look at Flip Solver in H11. All particles origin in POPs and after that are send to the solver. I tend to assume this alone makes Houdini's life harder in multi-million particles simulations.

The magic here is pure point-based environment. Such homogeneous world allows for many optimization. Everything is a point or a spring force ;).

Share this post


Link to post
Share on other sites

ok Symek...eagerly awating for your Houdini version of this tool :P how long will it take, a week, a week and a half? :P

Share this post


Link to post
Share on other sites

ok Symek...eagerly awating for your Houdini version of this tool :P how long will it take, a week, a week and a half? :P

Really? He's gonna do that tool for us? Wow, Symek! Such a kind soul! Take two weeks if you must!

Share this post


Link to post
Share on other sites

I suspect the biggest bottleneck of DOPs are POPs. Look at Flip Solver in H11. All particles origin in POPs and after that are send to the solver. I tend to assume this alone makes Houdini's life harder in multi-million particles simulations.

The magic here is pure point-based environment. Such homogeneous world allows for many optimization. Everything is a point or a spring force ;).

Well I said that you probably will have bottlenecks, but just guessing I dont know DOP too much to be sure about that.

But I am sure you can make fluid solver complitelly in SOPs where you sacrify flexibility to get total control of data management and get speed.

And talking about the point based solvers, is interesting how things are evolving, all the new stuff from maya is based on particle solvers, ncloth, nparticles, this amazing solvers from thiago the same, seems technology i tuning towards particle based solvers ...

Ans yes, still not having multithreading in POP for H11 is something I am really missing in the new release.

I would be happy with a new release if it will just add multithreading POPs.

Share this post


Link to post
Share on other sites
And talking about the point based solvers, is interesting how things are evolving, all the new stuff from maya is based on particle solvers, ncloth, nparticles, this amazing solvers from thiago the same, seems technology i tuning towards particle based solvers ...

I say we just skip particles and go directly to Superstrings :)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×