ZIH-Colloquium 2026

Table of contents

1. 1. History: ZIH colloquia of the past years

The ZIH colloquium is a public event and takes place usually the 4th Thursday of each month. Below you will find the current and former dates with all necessary information. You are cordially invited!

30 April 2026, 3:00 p.m., WIL A317 / Online BBB
Oliver Franzke (Double Fine Productions - Xbox Game Studios) - "Parallel Computing in Video Games"

Modern video games are among the most complex real-time parallel systems in consumer computing, combining SIMD, MIMD, and pipeline parallelism within a typical frame budget of 16.6ms on heterogeneous hardware. This talk begins with a brief look at how the exponential growth of output complexity across console generations - in graphics, audio, and CPU throughput - drove the transition from serial to massively parallel architectures. We then examine the domain-specific parallelism of the GPU and contrast its data-parallel strengths with gameplay logic, where circular dependencies, shared mutable state, and unpredictable memory access patterns pose fundamental challenges to parallelization. Using Unreal Engine as a practical case study, we explore the parallelization models and synchronization techniques a modern game engine employs to orchestrate sequential and parallel workloads across CPU threads and GPU under real-time constraints.

Oliver Franzke is a Principal Engineer at Double Fine Productions (Xbox Game Studios) with nearly 25 years of professional experience in the games industry. Across more than 20 shipped titles at studios including LucasArts, Sony, and Double Fine, he has worked on graphics, engine systems, and tools - from the Monkey Island Special Editions to Broken Age, for which he was Lead Programmer. Oliver holds a Diplom in Media Computer Science from TU Dresden, graduating with the highest distinction and receiving the Lohrmann Medal for outstanding academic achievement.

23 April 2026, 3:00 p.m., WIL A317/ Online BBB
Prof. Dr. Siegfried Raasch (Department of Meteorology and Climatology, IMUK, Leibniz University Hannover, Germany; Pecanode GmbH) - "The PALM Model System Release 25.10 - An LES Code for Basic and Environmental Research – Capabilities and Recent Applications" slides

Originally applied to study the convective atmospheric boundary layer (CBL), largeeddy
simulation (LES) is now used in many areas of science. This is mainly due to a
substantial increase in available computing resources. State-of-the-art massively
parallel computers have opened the field to a wide variety of new applications. On
these machines, simulations with extremely large numerical grids - up to 40003 grid
points and beyond - are currently performed within acceptable timeframes. In
Meteorology, besides for the fundamental research on neutral and stable stratified
flows, where the typical eddy size is much smaller than in purely convectively driven
flows, LES is increasingly used for more applied topics such as air-pollution
modeling, flow around buildings, wind energy, and aircraft operation.

The LES model PALM (PArallelized LES Model), originally developed at the Leibniz
University Hannover (LUH), has become a community model over the last 10-15
years. Besides LUH, a team of national and international developers has significally
extended the original code with a wide range of features, including cloud physics,
chemistry, radiation, Lagrangian particles, and various nesting capabilities. A strong
focus is placed on the explicit representation of detailed surface processes, e.g.
resolved-scale vegetation and the complex geometry of buildings. PALM is well
optimized for state-of-the-art cache-based and vector processors and scales
efficiently up to 40.000 cores and beyond. The code is currently ported to run on
GPUs.

The talk will start with a short general introduction to atmospheric turbulence and
LES, and then provide an overview of PALM’s features, illustrated by results from
recent studies at IMUK.

Siegfried Raasch was a professor of theoretical meteorology at Leibniz University
Hannover until his retirement in 2024. He is now an employed by Pecanode GmbH,
where he continues to develop and maintain the PALM model, which he initially
began developing in 1997. His research interests focus on atmospheric boundary
layer turbulence, large-eddy simulation, and the optimization of numerical algorithms.

25 March 2026, 3:00 p.m., N63 A.103 / Online BBB
Dr. Jesse Veenvliet (Stembryogenesis Lab, Max Planck Institute of Molecular Cell Biology and Genetics, Dresden; Cluster of Excellence Physics of Life, TU Dresden; Center for Systems Biology Dresden) - „Building towards understanding — leveraging stem-cell-based embryo models to elucidate principles of development“

Our lab investigates the principles underlying robust formation of the mammalian body axes, which organize the future body plan. While gene-regulatory networks are well characterized, how micro-environmental inputs arising from the physiological micro-environment contribute to the robustness of axial patterning and morphogenesis remains unclear. This is largely due to the complexity and constraints of in vivo development. We use stem-cell-based embryo models, such as gastruloids and trunk-like structures, which combine scalability and ease of measurement and manipulation with in vivo-like patterning and morphogenesis, making them ideal to obtain quantitative, causal, mechanistic insights. We integrate embryo models with live imaging, omics, genetics, biophysics, and data science to uncover how interactions between the embryo model and its micro-environment drive robust morphogenesis.

Current research focuses on: (1) biophysical principles of axis elongation; (2) metabolic regulation of patterning and morphogenesis; (3) roles of extracellular matrix remodeling; (4) robustness in axial progenitor fate decisions; and (5) leveraging variation to predict and control embryo model behavior.

Jesse Veenvliet studied Medicine at the Radboud University (Nijmegen), followed by an MSc in Experimental & Clinical Neuroscience at Utrecht University. He then did his PhD in the field of molecular and developmental neuroscience in the lab of Marten Smidt, at the Rudolf Magnus Institute in Utrecht and the Swammerdam Institute of Life Sciences in Amsterdam. For his postdoc, he joined the lab of Bernhard Herrmann at the Max Planck Institute for Molecular Genetics in Berlin, where he developed one of the first stem-cell-based models of early mammalian development. In May 2021 he moved to Dresden to start his independent lab at the Max Planck Institute of Molecular Cell Biology and Genetics. His group develops and uses mouse and human stem-cell-based embryo models to elucidate the principles governing robustness of mammalian body plan formation.

History: ZIH colloquia of the past years

History 2025
History 2024
History 2023  
History 2022
History 2021
History 2019/2020
History 2018
History 2017
History 2016
History 2014/2015
History 2013
History 2012
History 2011
History 2010
History 2009
History 2008
History 1998 - 2007