High Performance Computing for EUR research – the slides and the pictures
On November 24th 2015 a fair amount of EUR faculty convened in the Polak Building to receive an introduction on high performance computing (hpc) services for EUR research. After the presentation (see for the slides here) a hands on session followed. Everyone received a demo account and actually connected to the LISA platform.
The SURFsara high performance computing services were presented by Jeroen Engelberts and Jeroen provided support during the hands on session, together with his SURFsara colleagues: Bas van der Vlies, Jaap Dijkshoorn and Maarten Kooyman. These are the specialists providing their support via firstname.lastname@example.org or face to face, during training sessions.
Matthijs van Otegem (director University Library) explained in his welcome speech that research support at the EUR is provided in partnership with USC IT and indicated that hpc services were identified as a growing need within the EUR.
EUR faculty have been using hpc services at SURFsara for some years, based on individual proposals at NWO for access to hpc services. But we learned that on a practical level additional support needed to be made available. This is now arranged in the SURFsara / EUR contract in which dedicated access to LISA and Cartesius is provided (job priority or: lesser job queues).
See also the case of Genoeconomics and their experience with the SURFsara hpc services:
We perform research at the intersection of molecular genetics and microeconomics. Our goal is threefold: First, to identify specific genetic factors that are responsible for the observed heritability of many economic preferences and outcomes. Second, to identify the many ways in which individual behavior and social institutions moderate or amplify genetic differences. And third, incorporating genetics into economic analysis to help economists identify and measure important causal pathways (which may or may not be genetic). Our research in this field has until now focused on educational attainment, subjective well-being, and self-employment.
Why do we need Lisa and Cartesius?
Genoeconomics research requires analysis of genetic datasets which are to big to analyze on a normal desktop PC. Our group has used Lisa for several years, but due to recent releases of large genetic datasets the storage capacity and analysis capacity of Lisa became unsufficient for specific analyses. Therefore, many of our analyses will now be moved to Cartesius. We have very good experiences with SURFsara: The staff is very responsive to questions and has been very helpful is thinking about practical solutions.
Niels Rietveld, assistant professor at the the department of Applied Economics, Erasmus School of Economics.
Since the SURFsara hpc session two EUR applications to LISA / Cartesius have already been submitted, granted and set up.
If you require the hpc services yourself or wish to gain experience in this field, please feel free to acquire your access to LISA or Cartesius at: https://userinfo.surfsara.nl/systems/lisa/account (see: Affiliates of the EUR).
See for SURFsara: https://www.surf.nl/en/about-surf/subsidiaries/surfsara
For Lisa (Research Capacity Computing Service (RCCS)): https://www.surf.nl/en/services-and-products/research-capacity-computing-service/index.html
Photographer: Arie Kers, Erasmus University Rotterdam
Click here to see the pictures.