• Subscribe

How does a star form?

Speed read
  • Poorly understood turbulence influences the formation of stars in our galaxy
  • New high-resolution simulation of interstellar turbulence includes role of magnetic fields & reproduces complex internal structure of real star formation regions
  • Visualization brings us closer to understanding how stars—and Earth—came to be

Human imagination has long been piqued by images of space. As telescopes, low-frequency arrays, and other technologies evolve, humanity is brought ever closer to vibrant, exotic, and mysterious phenomena happening far from our home solar system.

Expanding the scope of astrophysics. This visualization of the role of magnetic fields and interstellar turbulence in star formation is a finalist for Best Scientific Visualization at SC19 in Denver, Colorado (Nov. 17-22). Courtesy LRZ.

For astrophysicists and cosmologists, new technology enables far more than just seeing beautiful images of far-off realms. Technology plays an indispensable role in learning how stars are born, how they die, and how those massive explosions spread elements across the universe, ultimately bringing us closer to understanding how the Earth as we know it came to be.

In the past decade, high-performance computing (HPC) has further accelerated and expanded the scope of astrophysics, enabling increasingly larger, more detailed simulations of supernovae and galaxy formation, among countless other research areas.

<strong>Interstellar medium</strong> is a gaseous environment that fills the space between star systems, transporting elements across the universe and influencing the formation of new stars. Courtesy NASA, ESA, N. Smith (U. California, Berkeley) et al. and The Hubble Heritage Team (STScI/AURA). Recently, a team of researchers led by Australian National University (ANU) Professor Christoph Federrath used HPC resources at the Leibniz Supercomputing Centre (LRZ) in Germany to run the largest magneto-hydrodynamic simulation of astrophysical turbulence ever performed to include magnetic fields.

The team partnered with visualizations experts at LRZ and Intel to turn their massive calculation into a stunning visualization. The visualization, executed using the Intel OSPRay rendering engine and VisIt software, is a finalist for the Best Scientific Visualization & Data Analytics Showcase award at SC19 in Denver, Colorado.

Turbulent times

To the naked eye, the space between star systems like our own solar system seems to be a vast, featureless expanse, but this space is in reality filled by the interstellar medium (ISM). ISM is a gaseous environment that plays a major role in transporting elements across the universe, ultimately shaping galaxies by influencing the formation of new stars.

<strong>Turbulence is especially demanding</strong> to simulate, as there is no symmetry or geometric construct that the physicist can take advantage of. Courtesy LRZ.Much of the energy traversing the ISM was ejected from violent supernovae and stellar winds, leading to chaotic, turbulent motions. Turbulence is one of the last major unsolved problems in physics, and understanding turbulent motions on such a massive scale requires large-scale computing resources. 

“The general challenge with astrophysics simulations is that they are all multiscale problems,” said Dr. Salvatore Cielo, researcher at LRZ and leader of the visualization effort. “In order to understand star formation, we need to model objects often as large as galaxies, but with a resolution orders of magnitudes higher to properly account for the physics of the ‘building blocks’ of stars, such as turbulence.”

Cielo continues, “Turbulence is especially demanding to simulate, as it tends not to be localized, but rather volume-filling; there is no symmetry or geometric construct the physicist can take advantage of.”

The simulation, recently produced by Federrath, reached a resolution of more than 10,000 grids per each spatial dimension, for a total of more than 1,000 billion resolution elements— the largest such simulation ever performed. This was necessary in order to capture the transition scale between supersonic (or self-sustaining via shocks) and subsonic (more rapidly decaying) turbulence.

<strong>Simulating more than 1,000 billion</strong> resolution elements required more than 23 terabytes of disk space and used nearly all of SuperMUC-NG’s 6,336 nodes. Courtesy LRZ.Besides being the first to ever measure this scale, the simulation reproduces very accurately the complexity of the internal structure of real star formation regions, as shown in the visualization.

Each snapshot required more than 23 terabytes of disk space, creating an enormous amount of data to visualize. Using the OSPRay engine and VisIt, the team was able to take advantage of nearly all of SuperMUC-NG’s 6,336 nodes. The team used ray tracing—a more realistic and, in turn, more demanding method of rendering graphics that lends itself well to being parallelized on HPC resources.

This new and novel method allows for direct access to the compute power of SuperMUC-NG for the calculations, visualization, and post-processing rather than transferring data to dedicated systems equipped with a traditional graphical environment.

Read more:

Read the original article on LRZ's site.

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2019 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.

Republish

We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.