Search

Navigation and service

LRZ: Summer of Simulation 2017

For the second time, LRZ offered a special training programme, tailored for master and PhD students of the field of life sciences, to employ molecular dynamics or quantumchemical simulations to tackle their problems on current supercomputers. In the framework of its "Summer of Simulation" (SoS) programme, high-performance computing (HPC) experts and tutors of the LRZ Bio Application Lab (BioLab) supported nine selected applicants to port their problems onto SuperMUC. In sum, the paticipants were able to consume more than 60 million core hours of computing time on the LRZ high-performance computing (HPC) system. In late October, all course participants presented their projects and the progress they had made over the summer at a "SoS Results Workshop".

begin
08.Jun.2017
end
31.Oct.2017
venue
Garching bei München (LRZ)

Summer of Simulation 2017, LRZParticipants and tutors of the Sommer of Simulation programme 2017.
Copyright: LRZ

How can computer simulations help to develop a nose spray against MRSA germs? How do new photochemical materials split water for solar-energy production? How can theoretical chemistry help to neutralize mustard-gas remains of World War II lying on the bottom of the Baltic Sea? Those and other questions have been addressed during the second “Summer of Simulation” training, conducted by the BioLab of the LRZ application support. The project was initiated in 2016 to help Master and PhD students employing molecular dynamics or quantumchemical simulations to tackle their problems on current supercomputers. After submitting a short proposal in June 2017, nine participants were invited to port their problems onto SuperMUC. Supported by high-performance computing (HPC) experts and tutors from the LRZ BioLab, they optimized their setups and conducted scaling tests, which were included in individual HPC project proposals. After passing the scientific review, the participants could consume altogether more than 60 Mio core hours that they had requested.

Summer of Simulation 2017, LRZV2C cave presentation, participaants of the Summer of Simulation programme 2017
Copyright: LRZ

In late October 2017, all participants met for a results workshop at the LRZ and presented their projects and the progress they made over the summer. Noteworthy is that two projects originated in groups that are new to molecular simulations: J. Hermann from the group of Prof. Weuster-Botz (Institute of Biochemical Engineering, TU Munich) examines how mutations can influence the crystallization behaviour of proteins. From the group of Prof. Winter (Pharmaceutical Technology and Biopharmaceutics, LMU Munich) S. Eisele and A. Tosstorff investigate how detergents stabilize pharmacological proteins against aggregation and thermal denaturation. Here, MD simulations help them to interpret their own experimental findings.

Further biochemical projects from the “Summer of Simulation” dealt with the parametrization of efficient implicit solvent models (M. Zachmann, group Prof. Antes, TU Munich) or how chemical modifications influence the switching behaviour of GTPase signalling proteins (D. Pourafar-Dehkordi, group Prof. Zacharias TU Munich). Material science oriented projects dealt with the deposition of photovoltaic layers of surfaces (H. Mirhosseini, group of Prof. Felser, MPG Dresden), doping of metal-oxides to improve their piezoelectric properties (M. Falkowski, group of Prof. Kersch, Munich UAS), and nanoparticles for electrolytic cell membranes (J. Timmermann, group of Prof. Reuter, TU Munich). For the latter project, help with visualization was also given by Ruben Garćıa Herńandez (V2C group of the LRZ) within the framework of the NOMAD project (Novel Materials Discovery, www.nomad-coe.eu). During the workshop, he also showed a presentation of the 3D visualization cave (fig. 2) and VR headsets for visualizing molecular structures and dynamics.

Summer of Simulation 2017, LRZCarbonitride systems for photochemical water splitting.
Copyright: J. Ehrmeier (TU Munich)

J. Ehrmaier (group Prof. Domcke, TU Munich) investigated the optical properties of carbo-nitride molecules of increasing complexity in solution, see Fig. 3. This class of molecules can split water upon photon absorption and thereby produce hydrogen gas. Understanding these processes and selecting optimal molecules could provide the technology for photochemical solar-energy production. Finally, T. Kl̈offel (group Prof. Meyer, FAU Erlangen) investigated how sulphur mustard is adsorbed and hydrolysed on zinc-oxide surfaces. More than 7,500 tons of this chemical warfare agent out of German stocks have been dumped in the Baltic Sea after World War II. As the grenades are starting to deteriorate, they pose a major hazard for the ecosystem and the neighbouring countries. Experiments with these substances are difficult and expensive due to the high safety level required. Here, computer experiments are a viable and risk free alternative. In a recent KONWIHR project, T. Kl̈offel optimized the code path within CPMD needed for these computations. With the improved node-level and parallel performance this system now runs one order of magnitude faster, which now makes the chemical reactions accessible in the simulation.

Summer of Simulation 2017, LRZSulfur mustard adsorbed on a zincoxide surface.
Copyright: T. Klöffel (FAU Erlangen-Nürnberg)

In January 2018, the participants will submit their final reports. Due to the great success and positive feedback from the participants, the LRZ will prepare the next instance of the “Summer of Simulation”, hopefully with new exiting projects to come.

Acknowledgments: We are indebted to the SuperMUC steering committee, particularly Prof. Wellein for his support and the referees for their fast review of the projects.

— Gerald Mathias, Momme Allalen

(This report was first published in Quartl, Issue # 84, 4/2017. Quartl ist das offizielle Mitteilungsblatt des Kompetenznetzwerks für Technisch-Wissenschaftliches Hoch- und Höchstleistungsrechnen in Bayern [KONWIHR] und der Bavarian Graduate School of Computational Engineering [BGCE].)