BlaBlubbBlog - Wissenschaft

Groundwork

nospam@example.com (Steffen) — Fri, 22 Aug 2008 18:46:00 +0200

And so it begins...

Visualization II

nospam@example.com (Steffen) — Wed, 28 May 2008 08:00:00 +0200

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.
Continue reading "Visualization II"

Visualization I

nospam@example.com (Steffen) — Thu, 22 Nov 2007 17:39:00 +0100

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 512Â³ 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 800Â³ grid cells. The actually rendering was quite fast, 5 minutes on a normal P4 desktop machine.

Number crunching

nospam@example.com (Steffen) — Fri, 01 Dec 2006 00:01:00 +0100

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:

Position of all halos with more than 10k particles, colour coded by the processor number which analysed them.

Continue reading "Number crunching"

Spot the mistake

nospam@example.com (Steffen) — Thu, 16 Nov 2006 19:13:33 +0100

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?

Click for larger

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?

Continue reading "Spot the mistake"

Hubble: Fifth Servicing Mission (SM4)

nospam@example.com (Steffen) — Thu, 02 Nov 2006 21:24:31 +0100

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.
Continue reading "Hubble: Fifth Servicing Mission (SM4)"

Planets. Again.

nospam@example.com (Steffen) — Thu, 24 Aug 2006 19:18:00 +0200

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

Continue reading "Planets. Again."

Die Planeten

nospam@example.com (Steffen) — Thu, 17 Aug 2006 18:55:00 +0200

Es gibt etwas neues von unseren n Planeten, das Planet Definition Committee hat eine neue Definition erarbeitet (siehe Pressemitteilung). Wenn diese Definition tatsÃ¤chlich so ratifiziert wird (offiziell ist es nÃ¤mlich noch nicht), dann haben wir mit einem Schlag statt neun nun zwÃ¶lf Planeten; und dabei wird es wohl auch nicht bleiben.

Die neuen Mitglieder in der Planetenfamilie wÃ¤ren dann Ceres, Plutos Mond Charon und 2003 UB313 (alias Xena, wir erinnern uns). Ãœber den endgÃ¼ltigen Namen von UB313 wird wohl wÃ¤hrend der noch stattfindenden General Assembly nicht mehr entschieden, aber wenn man einfach hartnÃ¤ckig Xena verwendet, hilft es vielleicht. Eine hÃ¼bsche Ãœbersicht Ã¼ber das 'neue' Sonnensystem findet sich hier.

Ein wenig verwundert bin ich Ã¼ber die Aufnahme von Charon, aber die BegrÃ¼ndung ist halbwegs Ã¼berzeugend: Bei Pluto und Charon handelt es sich um einen Doppelplaneten. Um ein Doppelplanet zu sein, mÃ¼ssen beide Objekte fÃ¼r sich genommen die Planetendefinition (massiv genug, damit sich eine sphÃ¤rische Form einstellt) erfÃ¼llen und zusÃ¤tzlich muÃŸ das Baryzentrum des Doppelsystem auÃŸerhalb der Objekte liegen (unser Mond ist demnach kein Planet, auch wenn er fÃ¼r sich genommen massiv genug wÃ¤re, aber das Baryzentrum des Erde-Mond-Systems liegt innerhalb der Erde).

Im weiten Netz kochen natÃ¼rlich schon die Diskussionen Ã¼ber Sinn und Unsinn dieser Definition hoch, siehe zum Beispiel hier (Cosmic Variance). Schauen wir erstmal, ob die Definition so abgesegnet wird. So ganz sicher bin ich mir da nÃ¤mlich nicht. Dann kann man auch sagen, ob meine Prognose richtig war.

Erfreuliches

nospam@example.com (Steffen) — Sat, 01 Jul 2006 00:28:35 +0200

gibt es von Hubble zu vermelden: die ACS geht wieder. Phew. Freut nicht nur die, die Beobachtungszeit haben und nun hoffentlich spannende Ergebnisse erzielen kÃ¶nnen, sondern auch die Freunde der bunten Bilder (noch mehr).

Nix und Hydra

nospam@example.com (Steffen) — Sun, 25 Jun 2006 12:32:00 +0200

Es gibt mal wieder Zuwachs im Kreise des Sonnensystems: Zwei letztes Jahr mit Hubble entdeckte Plutomonde haben nun offiziell ihre Namen bekommen, eben Nix und Hydra. Hydra kennt jeder, Nix hat mich verwirrt. Gemeint ist Nyx, die (griechische) GÃ¶ttin der Nacht, oder auch fÃ¼r uns eher in der rÃ¶mischen Mythologie Bewanderte, Nox. Da aber schon ein Asteroid diesen Namen trÃ¤gt, wurde mit Nix das Ã¤gyptischen Pendant genommen. Und wo wir schon bei Hubble sind, wie Heise gerade berichtet, ist die 'Advanced Camera for Surveys' (ACS) ausgefallen. Hoffen wir mal, daÃŸ die Techniker das wieder hinbekommen, da eine erneute Service Mission (mit einem Shuttle hinfliegen und notwendige Arbeiten durchfÃ¼hren) ja wohl noch in den Sternen steht.