Just before Christmas, LUMI was ready to receive applications for Danish projects from the Danish universities and DeiC. This was an excellent opportunity for the researchers at SDU, whose needs for HPC resources were not satisfied via the national call, to apply for additional resources.
LUMI (Large Unified Modern Infrastructure) is a pan-European pre-exascale supercomputer, able to provide computing power of up to 552 petaflops. SDU and the other Danish universities participate in LUMI under DeiC.
The allocation period for the first call was aligned with the allocation period for the national HPC resources, i.e. from 1st of January to 31st of April.
As the GPU resources on LUMI will not be available until medio 2022, only CPU core hours were allocated with this call.
In total, 4 projects at SDU were allocated resources on LUMI for the first period:
Quantum Chemistry for all parts of the Periodic Table – from materials to proteins
PI: Associate Prof. Erik Donovan Hedegård, FKF
Co-PI: Prof. Hans Jørgen Aagaard Jensen, FKF
Short abstract: Transition metals are integral parts of materials and catalysts and about one third of all enzymes contain a transition metal atom. The HPC resources applied for are for quantum chemical calculations to gain deeper insight on transition metal chemistry.
Into the QCD phase diagram
PI: Associate Prof. Benjamin Jäger, IMADA
Short abstract: The strong force (QCD) has a rich phase structure at finite temperature and chemical potential. We are tackling the sign problem at small quark masses using the state-of-the-art algorithmic and simulation techniques based on the openQCD software.
Annexin-mediated membrane repair in cancer
PI: Associate Prof. Himanshu Khandelia, FKF
Short abstract: In this project, we will attempt to disrupt plasma membrane (PM) repair in cancer cells, and reverse engineer non-repair proteins to repair cancer cells and vice-versa. Cancer cells PM are under intense rupture stress due to increased metabolic and cell division rates, and increased membrane dynamics resulting from invasive behaviour through the dense extracellular matrix. To maintain membrane integrity, cancer cells overexpress plasma membrane repair machinery, which all eukaryotic cells employ to cope with membrane disruptions. Members of the annexin (ANXA) protein family are Ca+2-sensitive phospholipid-binding proteins that initiate and regulate membrane repair. Despite the essential function of membrane repair in maintaining cell survival, PM repair mechanisms in healthy and diseased cells remain poorly understood. Using large-scale molecular simulations on HPC systems, we will elucidate fundamental biophysical mechanisms of PM repair with particular focus on the disruption of these repair mechanisms in cancer cells. We expect to improve our understanding of how ANXAs repair membranes lesions, and how such repair mechanisms can be compromised in cancer cells using small molecules that inhibit ANXA mediated repair, potentially leading to unique anticancer treatments. The project is a close collaboration with a funded consortium comprising of three other experimental groups in biophysics and cancer biology in Denmark.
Black Hole Image reconstruction for the Event Horizon Telescope via Bayesian parameter estimation
PI: Assistant Prof. Roman Gold, FKF
Short abstract: The computer resources support analyses within the Event Horizon Telescope collaboration including image reconstructions of the Black Holes in M87* and Sgr A*.
