Wissenschaft

And so it begins...

So I have been playing around with different visualization tools the last few days. Generally, I use gnuplot for all my plotting purposes and it does a great job producing paper grade plots. It even works quite well for just having a quick look at 3D data (like particle distributions et al.) or also 2D scalar fields, like a density field for example.

However, doing really fancy 3D visualization is not its purpose. I have been testing povray for that before and I got a few fancy images. But this does not quite give this immediate feedback of what you are doing: changing the colour mapping requires, naturally, a re-render, which for a decently sized volume data can take a couple of minutes.

I always wanted to look into visualizing all my fancy simulations. Never had the time, nor the motivation to do so. Until last night, at least with respect to the motivation, time I still don't have, but, oh well... I started to read up on how POV-Ray supports volumetric density media. Paul Bourke has a very nice article on that topic, which basically explains everything I needed. So the only thing that I need to do is to read all the particles from the simulation snapshot and put them on a grid. In a first approximation I used a cloud in cell (CIC) prescription, which means that a particle gives it's mass only to the cell it is located in, to calculate the density field. Once that density field is converted to a suitable binary file all that it is left to do is to write a really small POV-Ray scene. I basically used the MRI example from Paul's page to construct my scene and have a first working colour map (the tricky part).

And 'lo and behold, that is what I got (click for larger pop-up):



Not too bad. I need to play around with the colour map though to highlight the structure better. But already in this rough first attempt you can nicely see the filamentary and clumpy structure of the matter distribution. It is good to have a potent machine to do that on though. The density field is generated from 512³ particles (that is, 134 and a bit million). The raw particle data is already 4GB, the resulting density field (16bit dynamical range) 1GB since I used 800³ grid cells. The actually rendering was quite fast, 5 minutes on a normal P4 desktop machine.

As mentioned further down, I have been running quite some debugging runs, trying to beat out the glitches in the code. I am working on getting a parallel version of our halo finder to function properly and because I had to shift a lot of global variables around it took a while to find the last wrongly set one. But finally it works and I have been burning some CPU cycles tonight:

Okay, picture below is supposed to look like a universe. It is showing the position of identified halos in a simulation box. Now, is it not a bit weird that everything is squeezed in one corner?


I stared at that for five minutes, trying to figure out what went wrong. Maybe plotting the wrong columns of the datafile (there is much more stored than just the positions of the halos)? Nope, these are the right columns. Plotting the wrong files?

As has been announced already 2 days ago, NASA decided on having a servicing mission to the Hubble space telescope. After the Columbia accident Hubble's future looked dark, as the much needed maintenance visit was stopped due the halted Shuttle programme and NASA even decided to give up on Hubble, leaving us with no optical space telescope. The main problem are the gyroscopes which are needed to do the 'steering'. The abandonment of the space telescope yielded an uproar in the scientific community, but with the Shuttles not flying and even if flying their restriction to deliver parts to the ISS put a service mission far off.

Sooo... troubles resolved, we finally have a definition of what a planet is. Sharp as a spoon. As the most visible effect, Pluto lost its status as a planet and also nothing else new got added to the elitary club of planets. Lets have a look at the definition:

(1) A "planet"¹ is a celestial body that (a) is in orbit around the Sun, (b) has sufficient mass for its self-gravity to overcome rigid body forces so that it assumes a hydrostatic equilibrium (nearly round) shape, and (c) has cleared the neighbourhood around its orbit.

(2) A "dwarf planet" is a celestial body that (a) is in orbit around the Sun, (b) has sufficient mass for its self-gravity to overcome rigid body forces so that it assumes a hydrostatic equilibrium (nearly round) shape² , (c) has not cleared the neighbourhood around its orbit, and (d) is not a satellite.

(3) All other objects³ except satellites orbiting the Sun shall be referred to collectively as "Small Solar-System Bodies".

¹The eight planets are: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune.
²An IAU process will be established to assign borderline objects into either dwarf planet and other categories.
³These currently include most of the Solar System asteroids, most Trans-Neptunian Objects (TNOs), comets, and other small bodies.

-- IAU Press Release