Does it Compute? - Conference

Thursday 30 Oct 2025, 08:00 → 19:00 Europe/Amsterdam

Main Auditorium (Feringa Building)

Brian Setz (Rijksuniversiteit Groningen), David Visscher (Rijksuniversiteit Groningen), Event Mailer, Fokke Dijkstra (Rijksuniversiteit Groningen), Max Roeleveld (Rijksuniversiteit Groningen)

A Conference for Public Sector Computing

At Does it Compute we bring together practitioners, researchers, and educators to discuss the application of compute resources in their daily activities. Expect talks from practitioners about how they provide compute, sessions by researchers on the way compute powers science, and presentations by educators on how they employ compute and teach with it.

Come and enjoy a day full of insights, perspectives, and technical depth.

08:30 → 09:00 Registration / Coffee

09:00 → 12:00 Morning Sessions: Keynotes

Convener: David Visscher (Rijksuniversiteit Groningen)

09:00 Opening and Welcome

An opening and welcome to Does it Compute?

Here we'll give an overview of the event, house rules, and schedule.

Speaker: David Visscher (Rijksuniversiteit Groningen)

09:15 Something in the water

Speaker: Wim Nap (Rijksuniversiteit Groningen)

10:00 Break

10:15 Computer Labs at Scale: Cattle, not Pets

Gone are the days of carefully maintained computer lab machines ("pets"); today's computational demands require computer labs that are elastic, cloud-native, infrastructure-as-code environments ("cattle"). This talk explores how the University of Groningen has transformed its traditional computer labs into modern on-premise cloud-first computer labs. We will have a closer look at the tools and services that the UG developed to make this transformation possible, and the role of HPC in this journey. Finally, we'll also look ahead at what is in store for our computer labs in the future.

Speaker: Brian Setz (Rijksuniversiteit Groningen)

11:00 Break

11:15 The future of HPC: Trends and upcoming challenges

At the European level, significant investments in HPC systems are continuing. This opens opportunities for addressing various grand challenges in science and society. In this talk we will review some selected trends and, in particular, discuss various challenges to further progress on providing HPC resources and services as well as to facilitate their efficient exploitation for relevant challenges. We will do this for different levels, including hardware architectures and technologies, HPC infrastructures, and mathematical methods and algorithms.

Speaker: Dirk Pleiter (University of Groningen)

12:00 → 13:00 Lunch break

13:00 → 16:00 In Practice: Afternoon Sessions

Deep dive into what it means to run compute facilities in practice.
Expect technical talks from experts in the field.

Convener: Steven Woudstra (Rijksuniversiteit Groningen)

13:00 From Complexity to Simplicity: Our Cloud Infrastructure Redesign

The University of Groningen recently celebrated 60 years of computing at the university. A lot has changed in that time, but a constant factor is running calculations and analysis on our own hardware. Currently, this involves two clusters in two data centers and several petabytes of storage, built entirely with open source technologies. With this, we provide 2300+ researchers with computing power that matches their specialized needs without the prohibitive costs of commercial clouds, enabling research ranging from medical science to space exploration.

We want to show you what’s behind the blinkenlights: discover how we build and manage the foundational infrastructure under our compute clusters. We’ll cover our philosophy, the difficulties along the way, and our ambitious new project to redesign our entire infrastructure from scratch.

Key terms: configuration management, OpenStack, Ceph, software defined networking

Speakers: David Visscher (Rijksuniversiteit Groningen), Max Roeleveld

13:45 Break

14:00 I Can't Believe It's Not a VM! systemd-nspawn Containers Demystified

What if you could launch a full OS environment—with systemd, networking, persistent storage—in and full package management in milliseconds, not minutes? systemd-nspawn containers blur the line between containers and VMs, offering a drop-in, near-native performance without hypervisors.

In this talk, I’ll show you how to:
- Howto build and deploy VM-like nspawn containers
- Replace resource-heavy VMs with drop-in lightweight nspwan containers
- Boot a fully isolated Debian/RedHat/Arch systems ultra small and ultra fast
- Deploy nspawn containers seamlessly alongside your VM's

Walk away ready to deploy your first "VM-like" container before the next coffee break.

Speaker: Onno Ebbinge

14:45 Break

15:00 CTA: CERN’s Open-Source Tape Software for Archiving at Scale

CERN (European Organization for Nuclear Research) is home to the world's largest particle accelerator (Large Hadron Collider, LHC) that produces massive amounts of data each year. CERN's Storage and Data Management Group is responsible for enabling data storage and access for the CERN laboratory, in particular the long-term archival, preservation and distribution of LHC data to a worldwide scientific community (WLCG).

The CERN Tape Archive (CTA) software manages more than an exabyte of data across 7 tape libraries and roughly 70.000 tapes. To sustain write throughputs of tens of gigabytes per second, a flash-based disk buffer sits in front of CTA, allowing tape drives to write at near peak efficiency.

This high-efficiency archival service runs on open-source software and is deployed on-premises using commodity hardware. This talk will give a high-level overview of CTA, its deployment at CERN and the various design principles enabling its high performance.

Speaker: Niels Bügel (CERN)

15:45 Break

13:00 → 16:00 in Research: Afternoon Sessions

Sessions on the application of compute in research.
Expect talks from researchers on the way the employ compute to solve problems and do science.

Convener: Pedro Santos Neves (University of Groningen)

13:00 Updates on the Hábrók Compute Cluster

Hábrók, the HPC cluster of the University of Groningen, has been operational for two years. In this talk I will discuss some highlights and challenges during these year and discuss future plans.

Speaker: Fokke Dijkstra (Rijksuniversiteit Groningen)

13:25 From molecules to mechanisms: studying protein aggregation prevention with Hábrók

In this talk, I will share how large-scale molecular dynamics simulations are used to understand the behaviour of disease-related proteins and their regulation by molecular chaperones. This research, situated at the intersection of structural biology and computational science, relies heavily on the university's high-performance computing infrastructure, particularly Hábrók and Snellius. I will discuss how access to publicly funded compute and storage resources enables us to simulate biologically relevant systems over microsecond timescales, which would otherwise be unfeasible. I will also reflect on the practical aspects of working with these systems, ranging from managing long-running jobs and storage demands to optimizing simulations for specific hardware architectures.

Speaker: Vasista Adupa (Faculty of science & engineering, University of Groningen)

13:45 Break

14:00 An overview of EESSI

The HPC team of the University of Groningen is one of the main founders of the European Environment for Scientific Software Installations (EESSI). This collaborative project focuses on providing a common scientific software stack that can be used on virtually any Linux machine (and Windows, using WSL), including workstations, laptops, virtual machines, and HPC clusters.

By having this uniform software stack available, it will be much simpler to hop between different systems. You can first try to run the software on your own computer, then run the exact same version on for instance an HPC cluster like the university's Hábrók cluster, and even do the same on larger national or European clusters.

In this presentation we will give an overview of EESSI. We will also do a live demo to show how easy it is to use EESSI on the Hábrók cluster and on your own computer.

Speaker: Bob Dröge (Rijksuniversiteit Groningen)

14:25 Leveraging High-Performance Computing in Analytical Sociology

The rapid growth in large scale online data availability and the advancement of computational techniques have transformed the quantitative approach to social sciences. In this talk, I will explore how leveraging the Habrok cluster has empowered us to tackle complex sociological questions that were previously challenging due to methodological bottlenecks. Highlighting specific cases, such as imputing missing network data and employing large-scale NLP on over millions social media comments, I will illustrate how computational capabilities are becoming vital for analytical sociological research.

Speaker: Daniel Cowen (Faculty of social sciences, University of Groningen)

14:45 Break

15:00 Putting deep learning segmentation models in radiotherapy to the test

At the UMCG radiotherapy department, cancer patients are treated with targeted radiation. To plan treatment, clinicians outline tumours and nearby healthy organs on CT scans. This step is crucial but time-consuming. Therefore, deep learning segmentation models are now used in the clinical workflow. However, their outputs still require human oversight, as accuracy varies between cases. The resulting time-gain of automated segmentation is therefore limited. To address this, we use the Hábrók computer cluster to explore two directions: (1) uncertainty quantification, enabling models to indicate their confidence and better guide human oversight, and (2) modelling the impact of local segmentation errors, since not all errors equally affect the final treatment plan and some may have lower priority for correction.

Speaker: Joëlle van Aalst (UMCG)

15:25 Discussion: The future of Hábrók

In this session we want to wrap-up the research track with a discussion on the future of compute clusters like Hábrók.
What kind of service(s) does the university need in the future for providing researchers with compute and storage?
We'll ask the audience to participate by responding to a couple of questions.

Speaker: Fokke Dijkstra (Rijksuniversiteit Groningen)

15:45 Break

13:00 → 16:00 In Education: Afternoon Sessions

Talks about the application of compute in education.
Expect talks from educators on how they apply compute in their programmes.

Convener: Brian Setz (Rijksuniversiteit Groningen)

13:00 Automating Assessment in an HPC cloud

A cornerstone of modern, large-scale education is the use of various automated assessment methods to alleviate staff and improve the learning experience for students. Computing Science is a field that lends itself particularly well to assessment automation, given the well-defined nature of typical programming assignments.

In recent years, rapid growth of the Computing Science programme in Groningen has accelerated the adoption of automated code assessment in a wide range of courses, posing unique challenges to assignment design and automation setups. In this talk, we will present some case studies of recent adopters, and take a look into the future for a perspective on how a modern HPC cloud enables large-scale, highly-parallel, cloud-native, yet secure automated assessment.

Speaker: Floris Westerman (RUG Digital Lab)

13:45 Break

14:00 Azure Labs

Speaker: Jos Bos (Hanzehogeschool Groningen)

14:45 Break

15:00 Bringing Datacenter Infrastructure into the Classroom

As technology has progressed, the datacenter has become more important for IT education than ever before. Not only has it become more important than ever for students to know how to work with a datacenter and deploy applications to it, but giving students access to datacenter machines solves problems such as needing to run high-compute tasks not suited for the students' laptops (e.g. artificial intelligence), avoids complex installations and debugging for students by providing them with pre-created environments, and allows for assignment setups such as environments for cybersecurity courses where students need to attack targets within a closed off network. Digital Lab has created Virtual Labs, a service that distributes datacenter resources to students for use in their courses.

Speaker: Lars Andringa (RUG Digital Lab)

15:45 Break

16:00 → 18:00 Borrel