In terms of computing power, we astrophysicists are insatiable
Astrophysicist Hans-Thomas Janka, a scientist at the Max-Planck-Institute for Astrophysics and lecturer at the Technical University of Munich (TUM), and his team have used SuperMUC excessively: A conversation about the benefits of supercomputers and the hunger for data in physics.
More than 570 million core hours on SuperMUC: Congratulations, you have been the heavy user of this high-performance computer for the last six years.
Professor Hans-Thomas Janka: For us, the SuperMUC was a gift. Previously, astrophysicists could only perform much smaller calculations and thus only calculate two-dimensional models. With the SuperMUC even three-dimensional simulations became possible - that was a huge breakthrough for us.
What are you researching with the help of supercomputers?
Janka: My team and I are working on the explosions of supernovae, massive stars and energy boosters in space. Every second a supernova explodes somewhere, in our galaxy only two to three of them can be seen as bright events with the telescope – per century! These explosions form neutron stars, black holes and chemical elements. Using data from supernovae, we study the formation of metals, matter and life, as well as the role of these stars in space.
Why does astrophysics create so much data?
Janka: We get a lot of data from stars and explosions from the cosmos by measuring radiation, elementary particles or gravitational waves. But we can only really observe a few developments. Therefore, astrophysicists develop models for evaluation and calculate them with the help of mathematical and physical equations. This easily produces terabytes of data that we can only analyze or visualize with high performance computers.
How can you be sure that the models and algorithms for evaluation are correct?
Janka: Astrophysicists need patience to confirm their theories. Certainly, predictions can be supported by comparing different and independent calculations, but certainty comes from observation and measurement. Models for the collision of two neutron stars, which I developed with colleagues in the 1990s, could only be reproduced in reality 20 years later.
Supercomputing and astrophysics - how is the partnership developing?
Janka: We are insatiable when it comes to data volumes and computing power. Today we can only calculate many models in two dimensions, and we would also like to refine the resolution of three-dimensional calculations from two to one degree - but this doubling would at least increase the computing time of today's computers eightfold.
And why do you need a computer centre like the LRZ?
Janka: Although we write most of the algorithms ourselves, the support for porting the data and applications or for visualization is enormously helpful. The LRZ has built up many competencies that we don't have. They will become even more important in the coming exascale and perhaps quantum computer times, because the machines and their programs will become even more complex.