Brug af supercomputing - Abacus 2.0, SDU

Supercomputerne – eller mere præcist High-Performance Computing - bruges i dag til at designe fly, skibe, ja selv tandbørster, og vejrudsigten ville være endnu mere uforudsigelig, hvis den ikke blev lavet af supercomputere. På den måde er High-Performance Computing blevet en del af vores hverdag uden, at vi nødvendigvis er klar over det. På Abacus 2.0 - Danmarks hurtigste og mest fleksible supercomputer- kan alle danske forskere og virksomheder bruge regnekraft til at skabe gennembrud inden for ingeniørkundskaben, grundforskningen og i erhvervslivet.

Abacus 2.0 – A state-of-the-art flexible and affordable solution optimized for a wide range of research from both the traditional HPC disciplines (natural sciences) and the emerging ones (social sciences)

High Performance Computing (HPC) previously the domain of theoretical scientists and computer and software developers is becoming ever more important as a research tool in many surprising areas. The use of HPC in modelling complex physical phenomena such as weather, fluid dynamics, molecular interactions, astronomical calculations and engineering design is well known to researchers in those fields.

HPC is also now being used in industry to improve products, reduce production costs and decrease the time it takes to develop new products.

As our ability to collect Big Data increases the need to be able to analyse the data also increases, this is an area HPC can be a most useful tool.

More recently HPC is being used by researchers in social media, semantics, geology, archaeology, materials, urban planning, graphics, genomics, brain imaging, economics, game design and even music. The list will continue to expand as more people are introduced to the possibilities of using HPC, bringing their own unique understanding of how it can be used in their fields.

User case 1:

Chemistry has seen a computational revolution within the last 15 years and theoretical chemists have provided computer programs which today serve as valuable research tools for predicting new materials by guiding synthesis and interpreting spectroscopy. Certainly, the field of computational chemistry profits from the enormous advancement of computer technology, but, much more, it benefits from the development of efficient algorithms and solution techniques of the relevant equations as well as of the derivation of suitable mathematical models for the description of chemical processes. Jacob Kongsted, professor at the Department of Physics, Chemistry and Pharmacy at SDU, has a mission to bridge the gap between the demand for theoretical support and the missing computational technology by developing reliable and efficient tools for the simulation of advanced molecular spectroscopy and photochemistry for complex materials.

User case 2:

David C. Jinkins, assistant professor at the Department of Economics at the Copenhagen Business School, has an ongoing project on the effect of academic mobility between departments and the spread of knowledge as measured by citations. Using a new panel data set linking academics to departments and citations, he develops and estimate a dynamic model of location choice in which an idea is more likely to be encountered when colleagues already know about it. Several exercises indicate that coworker knowledge significantly affects the probability of learning about a new idea. Counterfactual exercises show that labor mobility increases the speed at which new ideas spread between locations, makes locations more uniform in the fraction of people who know about a new idea, and raises the percentage of people who know about a new idea at a given time.

User case 3:

Esmaeil S. Nadimi is associate professor in signal processing at the Faculty of Engineering, SDU with core research competences in big data analysis of wind turbines and design of non-invasive mini-robots. He is a part of the Machine Learning Group that runs a project in collaboration with Siemens Wind Power (SWP) and Lindøe Offshore Research Centre (LORC). The aim of the project is to develop a computationally fast and inexpensive fault detection and prediction method for offshore wind turbines. The group has access to the database of all the turbines (onshore and offshore) commissioned by SWP for the last 20 years around the world.

User case 4:

Matteo Pilati, PhD Student in Archeology at the Aarhus University is a member of the HPC Archaeology Project. Abacus 2.0 is used to generate very detailed 3D documentation of series of images which allows to get a on the spot overview of an excavation and thus make informed decisions about the next steps to take. It also allows other scientists subsequently to study the findings. For more information see also: eScience åbner nye døre for arkæologien.