Millennium simulation in Celestia

[Cham]

I'm currently working on a new addon for Celestia : the millennium simulation, as shown here :

http://www.mpa-garching.mpg.de/millennium/

See for example this page (several animations) :
http://cosmicweb.uchicago.edu/filaments.html

Simple resume on Wikipedia :
http://en.wikipedia.org/wiki/Millennium_simulation

Here's a picture of the current prototype in Celestia (about 53 000 "galactic-particles" in the cube, so it's running very smoothly in Celestia) :

cube.jpg

I'll probably make a kind of "multi-layers" for Celestia, to simulate the animation shown here :
http://cosmicweb.uchicago.edu/images/mov/bnr_full2.mpg

My only problem for the moment is the choice of colors. I'm interested to use a color blending effect, to reproduce the colors on the pictures as seen here :

http://www.mpa-garching.mpg.de/galform/ ... illennium/

[t00fri]

See also my respective discussion with video material a long time ago, here:
http://forum.celestialmatters.org/viewtopic.php?t=25

Or from a recent meeting (DESY Theory Workshop 2008) at my lab:
(via Lactea Project!)
http://th-workshop2008.desy.de/sites/si ... iemand.pdf

with these great videos from the speaker's site
http://www.ucolick.org/~diemand/vl/

http://www.ucolick.org/~diemand/vl/movies.html


Fridger

[ElChristou]

I'm currently working on a new addon for Celestia : the millennium simulation, as shown here :...

Yep, I also think that meanwhile the engine of Celestia is modified to render such structure, such addon is really welcome!

[Cham]

I made 24 animation "frames". This is trully amazing, in Celestia, while moving inside the cube and watching the animation step by step (using a "multi-layers" script). You can clearly feel galaxies clusters rotating around each others, "nodes" evolving, etc. Here's a picture of the 10 first "frames" (the jpeg compression is terrible on this picture) :

cube.jpg

The 24 ASCII CMOD files are taking about 38 MB on the HD (average of 1.6 MB per frame).

[Cham]

Here's a preview of the simulation of the universe *today*. The colors are coming from the original source (truncated into 12 shades). As usual, click the thumbnail for a larger view :

box1.jpg

box2.jpg

box3.jpg

[Cham]

Two more views of the cube (size of cube *today* is 279 MLY) :

box4.jpg

box5.jpg

Thanks a lot to Gerard Lemson, from the Millennium Database project, to help me in using the database query.

[ElChristou]

Seems pretty cool!

[Cham]

The Millennium Simulation multi-layer is merging very well with the SLOAN model. Very nice in Celestia :

cubeA.jpg

cubeB.jpg

I think the Millennium addon is ready.

[ElChristou]

Strangely I don't know why but I thought your cube would be much larger than the SLOAN model...

[Cham]

No, that cube is just a small portion of the whole Millennium simulation. The data of the whole simulation is MUCH larger and would be a total mess in Celestia and the files would be MUCH larger on the hard disk! The cube I made has a size of about 280 MLY, which is already pretty large (ASCII CMOD files are of about 36 MB on disk). You can fly inside the cube easily, and the view is already spectacular.

EDIT : The size of my box is actually 1/8 of the whole simulation box, hence [tex](1/8)^3 = 0.2\%[/tex] of the total volume. The whole simulation box was using periodic conditions on its edges, which is very commonly done for this kind of research on a super-computer.

ElChristou, do you want me to send you a model for testing, by email ? I should be able to do it later today.

[Cham]

Hehehee !

Soon : galactic interaction and mergers in full 3D for Celestia !

http://ozone.obspm.fr/horizon/galmer/

Too bad the sprites aren't working correctly in Celestia

EDIT : Look at this one :

http://www.projet-horizon.fr/IMG/gif/pa ... 048.1c.gif

[selden]

Cham,

Are you planning to provide ScriptedOrbit functions for the stars of the interacting galaxies?

(It'd be nice if ScriptedOrbit functions could work for galaxies, too. )

[Cham]

Cham,

Are you planning to provide ScriptedOrbit functions for the stars of the interacting galaxies?

(It'd be nice if ScriptedOrbit functions could work for galaxies, too. )

No. The galactic collision models will just be a bunch of CMOD "frames", that could be turned ON/OFF with a "multi-layers" script. I should be able to achieve a step by step animation like for the Millennium simulation. Each frame is made of several thousands of "particles", defined by the simulation computations.

[Cham]

Here's the kind of stuff I'm interested to build in CMOD format. I should be able to make some models during the following weeks.

http://cosmicweb.uchicago.edu/images/gal/gal_z4_rot.mpg

Of course, the model will be defined as a multi-layer (with step by step animation of the structure formation).

[ElChristou]

...ElChristou, do you want me to send you a model for testing, by email ? I should be able to do it later today.

I fear 36 Mo by mail is not a good idea... You should definitively find a server soon!

[Cham]

...ElChristou, do you want me to send you a model for testing, by email ? I should be able to do it later today.

I fear 36 Mo by mail is not a good idea... You should definitively find a server soon!

Well, I can send just some "frames" (each one is about 1.6 MB, unzipped).

The whole addon is approximately 10 MB, once zipped.

[ElChristou]

...ElChristou, do you want me to send you a model for testing, by email ? I should be able to do it later today.

I fear 36 Mo by mail is not a good idea... You should definitively find a server soon!

Well, I can send just some "frames" (each one is about 1.6 MB, unzipped).

The whole addon is approximately 10 MB, once zipped.

Then what about 2 mails of 5Mo?

[Cham]

Note to self : nice simulation to reproduce in Celestia at

http://www.mpa-garching.mpg.de/Virgo/hubble.html

EDIT : Argh ! First test model failed. The small corner shown below is a 2.1 MB ASCII CMOD file (a small subset of the whole data), while the whole data is supposed to completely fill the cube . Their data is WAAAAAY too large ! A complete CMOD model would take several GB on disk !8O

[ElChristou]

That's means, adios addon?

[Cham]

That's means, adios addon?

No. I just have to select some subsets. I'll have to experiment a lot with the huge data files. There are several simulations to play with, anyway: there are many independant simulation groups out there, which are giving access to their data. So many similar addons, with various approaches, using different values of the cosmological parameters (Omega ...), etc.

So the addon I gave to you by email is just a first one of this kind.

Instead of showing the galactic structure at a given epoch, using several frames to show the time evolution, I want to produce a pie shaped model of the simulated distribution of galaxies in space with the dependance on redshift, like the SLOAN model. See for example the picture here :

http://www.mpa-garching.mpg.de/Virgo/lc_lcdm.gif