Page 3 of 4

Posted: Sun Jan 15, 2006 8:01 pm
by Nick_Lambert
n_t wrote:Money_YaY!: I finally setup a mac here and tried to build a mac version, here it is (Python2.4): http://www10.informatik.uni-erlangen.de ... mac.tar.gz
I'm not sure how works together with the different OSX versions, though...
Many thanks Nils, I'll let you know how 10.4.4 works out. :D

Posted: Sun Jan 15, 2006 8:17 pm
by Nick_Lambert
Nick_Lambert wrote:
n_t wrote:Money_YaY!: I finally setup a mac here and tried to build a mac version, here it is (Python2.4): http://www10.informatik.uni-erlangen.de ... mac.tar.gz
I'm not sure how works together with the different OSX versions, though...
Many thanks Nils, I'll let you know how 10.4.4 works out. :D

Won't even launch :(

Posted: Mon Jan 16, 2006 4:42 pm
by n_t
hi Nick_Lambert, bad to hear - could you try to get the command line output (by starting Blender from a console or so)...? That sounds like the wrong python version, or some library problem.

Posted: Mon Jan 16, 2006 5:37 pm
by rcas
n_t wrote:hi Nick_Lambert, bad to hear - could you try to get the command line output (by starting Blender from a console or so)...? That sounds like the wrong python version, or some library problem.
If this was working fine (stable) in all platforms it would be a very good addition to the upcoming 2.41 release.

Posted: Mon Jan 16, 2006 11:43 pm
by LetterRip
rcas,

it is fairly unstable from my testing, and is far too late in the release cycle for a major addition - Ton would have kittens :)

2.42 or 2.50 whatever we call it is only a couple of months away, it isn't like it will be a half a year or so :)

LetterRip

Posted: Tue Jan 17, 2006 5:02 pm
by rcas
LetterRip wrote:rcas,

it is fairly unstable from my testing, and is far too late in the release cycle for a major addition - Ton would have kittens :)

2.42 or 2.50 whatever we call it is only a couple of months away, it isn't like it will be a half a year or so :)

LetterRip
I compile CVS myself, so I have no problems with that :)

I was just pointing out for the Stable version users, not crazy guys like me that use the CVS version, hehe.
But since it isn't that stable might aswell wait, yes.


Oh, I stand for 2.50, minor version should be for revisions or bug fixes.

-- Rui --

Posted: Tue Jan 17, 2006 10:21 pm
by simhar
hello,
is the lattice-boltzman-method, you are using in the implementation able to be used in professional cfd-simulations; i looked out and found just powerFLOW as an example for an professional cfd-solver based on lattice-boltzman (for example fluent is using finite-volume-method) ... so, do you think your implementation in blender is exact and stable enough to run scientific simulations or is it just "lookalike".

thanks
simon

Posted: Tue Jan 17, 2006 10:21 pm
by simhar
hello,
is the lattice-boltzman-method, you are using in the implementation, able to be used in professional cfd-simulations; i looked out and found just powerFLOW as an example for an professional cfd-solver based on lattice-boltzman (for example fluent is using finite-volume-method) ... so, do you think your implementation in blender is exact and stable enough to run scientific simulations or is it just "lookalike".

thanks
simon

Posted: Wed Jan 18, 2006 3:28 am
by STaRMaN
n_t, i had have to register only for congratulate you by this. It's great.
I'm making some tests, and i am :shock: too..

I hope that we have it in a next official release :)

if someone has some good tests (and/or ideas) could post them here..


Hello for all :)
(this is my first post)

Posted: Mon Jan 23, 2006 7:19 pm
by ZanQdo
Nils how much longuer do you think this would be in cvs? I'm using fluids for a project rigth now and I kind of miss moving obstacles-fluids :D

Posted: Mon Jan 23, 2006 9:28 pm
by LetterRip
ZanQdo,

it isn't in CVS yet, hopefully he will put it in once 2.41 comes out...

LetterRip

Posted: Tue Jan 24, 2006 9:08 am
by ZanQdo
Cool, this is a recent render of the thing I'm doing:

http://www.3developer.com/wip/FresaFluidWip.mov

but soon I will need a moving object scene, I could use a linux box but I'm also using node materials so I'm on a problem rigth now :P

Anyway thanks for everything

Posted: Mon Jan 30, 2006 3:51 pm
by n_t
simhar: the method was developed to do optimizations of metal foaming processes, so there's a physical background and it should approximate the underlying equations quite well. For any simulations with physical relevance I would recommend doing some accuray tests first, though... This powerFLOW software is the first commercial LBM program, but mostly because the method is quite new in contrast to other ones.

ZanQdo: good question, from the respones here I think at least the new API works quite well, so I hope I can start checking it in sometime soon. Nice animation btw.!

STaRMaN: if you have some nice animation with moving obstacles, you could also post it :)

Posted: Fri Feb 03, 2006 5:07 am
by ZanQdo
Hi nils, I have another question, is it going to be posible to make fluid simulation threaded?, we are using the solver quite a bit here but the baking times can get too long and I'm wasting a prosessor in my machine :?

EDIT: OK, installed a Linux box here, lets see how it works :D I think I can calculate the meshes in this test build and then load them in my recent cvs build for rendering with node materials, (that's a question, I jus't cant find the question mark in this Linux box LOL)

EDIT2: Answering my own question, yes I could load fluid mesh data easily in another proyect, cool :-)

Is it posible to make the inflow velocity Local? I want to rotate an inflow object to spread fluids in all directions but that does't work :P

Posted: Sat Feb 04, 2006 11:47 pm
by cekuhnen
very amazing looking movie.

how far will it be possible to letthe mesh look like natural water
while still being able to work with the application?

in the move the fluid looks very blotchy and you see the
underlying 3d mesh.

claas