Foto: Copyright HEPHY/Fotograf Gregor Schweinester
Computing power is becoming increasingly important, with the life sciences field being no exception. Frequently used methods such as DNA sequencing, theoretical modelling and scientific visualisations of structures typically require the support of powerful processors.
A High-Performance Computing (HPC) team of Gregor-Mendel-Institute (GMI), Institute of High Energy Physics (HEPHY), Institute of Molecular Biotechnology (IMBA), Stefan Meyer Institute (SMI) – all institutes of the Austrian Academy of Sciences, ÖAW – and the Research Institute of Molecular Pathology (IMP), joined forces at the Vienna BioCenter to establish the cloud infrastructure platform “CLIP”, a supercomputer cluster that improves computing speed manyfold. Thanks to CLIP, every scientist who uses so-called “in silico” methods – complex simulations –has the opportunity of a one-of-a-kind supercomputer at her or his fingertips – no matter the location of their laboratory. This includes research fields ranging from geology to physics, mathematics and even aerospace research.
Ronny Zimmermann, Head of IT at IMP, IMBA and GMI and master of the machine, is proud of his team’s achievement: “When using CLIP, scientists can dramatically reduce the time necessary for calculations. A task that usually takes many hours, can be completed in just a few seconds – this is comparable to several hundred workstations working together as one unit. We are very excited to report that as of October of this year, we already have over 400 users employing CLIP as part of their scientific studies”.
CLIP boasts impressive numbers, with more than 200 servers running on over 7700 CPU cores (regular computers have between four to six) and 112 GPU cards. 14 Austrian institutions are already currently using CLIP services, including numerous research groups from the Vienna BioCenter itself, and the Austrian Academy of Sciences.
“In less than a year, CLIP has become essential to our group for analyzing the large datasets from the LHC experiments. Not only does it meet the storage and I/O requirements, but the support crew also ensures a high degree of availability, and a foreseeable upgrade schedule facilitates the planning. To top it off, the GPU infrastructure is an ideal environment for cutting edge machine learning applications. All things considered, our new "home-base" outperforms most similarly sized competitors at the international level.” says HEPHY group leader Robert Schöfbeck enthusiastically.
CLIP now launched its own website, where scientists can easily access services and retrieve their data: https://www.clip.science/.
Quick facts about CLIP:
>7700 CPU Cores
112 GPU Cards (graphical cards, supports CPU and for scientific visualizations)
High-Mem Nodes with 1 and 4TB
I/O Nodes with 250TB storage and 170GByte/s read and write