September 16, 2026 - by CSCS
Name
Marina Krstic Marinkovic
Position
Assistant Professor in Computational Physics, Institute for Theoretical Physics, Department of Physics, ETH Zurich.
Area of research
Theoretical particle physics, high performance computing and quantum simulation.
My focus
In my research, I study nature’s fundamental interactions using large scale numerical simulations. I work on quantum field theories that describe the world of elementary particles and focus mainly on the theory of quarks and gluons, known as quantum chromodynamics (QCD). This theory cannot be solved with pen and paper methods alone, so we use supercomputers to simulate them on spacetime lattices, which is why the field is known as Lattice QCD. My group also develops new algorithms, including machine learning and quantum simulation methods, to study phenomena that are currently beyond the reach of conventional calculations, such as the dynamics of many-body quantum systems and matter under extreme conditions.
What supercomputing and Alps mean for me
Modern particle physics relies on large scale simulations that require computing power at the level of exaflops, making access to supercomputers such as Alps at CSCS essential. Alps allows us to perform simulations that would be impossible on smaller systems, both because of the sheer size of the calculations and because of the need for high precision calculations, requiring extremely large statistical samples, fine spacetime lattices, and millions of solves of large linear systems in order to control various systematic effects. We simulate quantum fields on four-dimensional spacetime lattices containing tens of millions of lattice sites and generate many data ensembles for different physical parameters that must be analysed. Access to Alps makes it possible to push these calculations to include effects of interplay between strong and electromagnetic interaction, test new algorithms, and explore questions that are directly connected to experiments.
What challenges do I face
Quantum theories describing fundamental particles are extremely complex and require both very large computational resources and the development of new algorithms. Many of the most interesting problems, such as the dynamics of quantum field theories, remain beyond the reach of standard simulation methods, which means that we must design new numerical techniques to study them reliably. Another challenge is the rapid evolution of supercomputing architectures. Modern machines rely on heterogeneous processors, GPUs, and highly parallel systems, and scientific codes must be continuously redesigned to run efficiently on these platforms. Achieving both high performance and the level of numerical precision required for modern particle physics calculations is therefore a constantly moving target. At the same time, these challenges often require close interdisciplinary collaboration between theoretical and experimental physicists, applied mathematicians, and computer scientists, which I find particularly rewarding and one of the aspects of my work that I enjoy very much.
What I like most about my work
What I enjoy most is working at the intersection of theoretical and experimental physics, and scientific computing. We use supercomputers to answer very fundamental questions about nature, and at the same time we develop new computational methods that can be useful far beyond particle physics. I also value the collaborative aspect of the field, where progress often comes from large international efforts combining theory, computation, and experiment.
My favourite recent project using Alps
One of my group’s recent projects on Alps, in collaboration with CSCS and supported by the Platform for Advanced Computing (PASC), focuses on accelerating lattice simulations of strong and electromagnetic interactions, needed, for example, for precision calculations of the muon’s magnetic moment. These simulations require solving the Dirac equation—a quantum-mechanical equation that is consistent with special relativity—millions of times on four-dimensional lattices and can only be performed efficiently on modern supercomputers such as Alps. In recent works, we developed optimized quantum chromodynamics and quantum electrodynamics (QCD+QED) measurement routines in the openQxD code for GPU architectures, enabling efficient measurements of QED corrections to observables of hadrons, such as protons and neutrons.
Career background
- 2009 – 2013 PhD in theoretical particle physics, HU Berlin, Germany
- 2012 – 2014 Research fellow / Postdoc, University of Southampton, UK
- 2014 – 2017 CERN Fellow, TH Department, Geneva, Switzerland
- 2016 – 2019 Hitachi Assistant Professor in High Performance Computing, Trinity College Dublin, Ireland
- 2020 – 2021 Junior Professor (W1), Arnold-Sommerfeld Center for Theoretical Physics, LMU Munich, Germany
- 2021 – Present Assistant Professor of Computational Physics (tenure-track), ETH Zurich