Eye catcher

HZDR Research Talks

The HZDR Research Talks usually take place on Wednesdays from 1:00 to 2:00 p.m.

The next dates are:


March 17, 2021

by Dr. René Hübner (Structural Analysis)


February 24, 2021

by Dr. Nina Elkina (IT - Infrastructure)

Regardless of a technical implementation of a reversible computer the erasure of one bit releases at least kBT log 2  ≈ 0.017 eV (at room temperature) of energy, as stated by Landauer's principle [1]. This makes the future of High Performance Computing (HPC) not only dark but also quite hot. And the slowdown in the computational power increasing rate is already reflected in the Top500 list of the most powerful computers in the world [2]. Therefore we have 2 questions on the agenda:
1) what is the next technology to catch up with constantly increasing needs for High Performance Computing and 2) how to survive the transition period while keeping the same pace of scientific progress. We are relatively lucky with the latter, since classical HPC has still a great room for improvement by exploiting the hardware and new numerical methods. In this regard having Hemera, the cluster on HZDR premises, allows to support current needs to predict near future demands of the local scientific community working with diverse computational models. I will briefly outline the computational landscape of HZDR, partially discovered during the first round of the seminar series 'Hardware and Numerics'. Then I will proceed to computationally intensive projects related to laboratory astrophysics and study of strong field QED effects in photon-photon collider experiment of LUXE [3] hosted at DESY. This experiment aims to measure quantum rates of electron-positron pairs production in substantially non-perturbative QED regime of field-assisted electron-positronpair creation and nonlinear Compton scattering. Due to multiple uncertainties in the field configuration the experiment needs to be supported with numerical simulations of realistic distributions of electromagnetic fields with subsequent optimization for a maximum yield of pairs. This requires solutions of transport equations either deterministic grid based or by means of stochastic Monte Carlo methods [4]. Both approaches are widely used for simulation in various scientific domains and searching for more accurate numerical implementations with fewer bit operations are of primary importance. The last but not least part of my talk will touch a problem of quantum computing, one of viable candidate to pick HPC's race in the post-silicon era. Most of quantum computing operations are unitary which implies reversibility. Therefor the restriction imposed by Landauer's principle will be mitigated. But hardware for quantum computers is still in its infancy and only a few qubits prototypes are commercially available. While attempts to simulate quantum computer algorithms on classical HPC face a problem of exponential complexity. From the other hand the data intensive nature of quantum computer simulators makes them a perfect testbed for benchmarking classical HPC clusters. I will conclude my talk with preliminary results of the Hemera quantum simulator project.

[1] R. Landauer, Irreversibility and Heat Generation in the Computing Process, IBM Journal of R&D, 5, 183-191 (1961)
[3] LUXE Collaboration Team, Conceptual Design Report for the LUXE Experiment, arXiv:2102.02032, (2021)
[4] Elkina N., Fedotov A., et al. QED cascades induced by circularly polarized laser elds, Phys. Rev. STAB 14, 054401 (2011).


The HDZR Research Talks are a program for HZDR employees of all research fields. Within this framework, scientists give lectures on current projects, methods and procedures and look forward to exchanging ideas with colleagues in order to jointly develop ideas and initiate projects. The Research Talks are also open to external interested parties.

  • Lectures and discussions are held in English.
  • Participation is free of charge.
  • Registration is not required. External visitors need to check in at the HZDR main entrance 30 min in advance.

Initiator:
Dr. Franziska Lederer (Helmholtz Institute Freiberg for Resource Technology, Dept. Biotechnology)
Phone: +49 351 260 2427
E-Mail: f.lederer@hzdr.de

Co-organizers:
Dr. Kristof Zarschler (Institute of Radiopharmaceutical Cancer Research, Dept. Radionuclide Theragnostics)
Phone: +49 351 260 3678
E-Mail: k.zarschler@hzdr.de

Dr. Peter Zahn (Institute of Ion Beam Physics and Materials Research, Dept. Scaling Phenomena)
Phone: +49 351 260 3121
E-Mail: p.zahn@hzdr.de

Dr. Björn Drobot (Institute of Resource Ecology, Dept. Biogeochemistry)
Phone: +49 351 260 2759
E-Mail: b.drobot@hzdr.de

Dr. Sandra Hamann (Research Programmes and International Projects)
Phone: +49 351 260 3649
E-Mail: s.hamann@hzdr.de


Past lectures: