|Name:||BoF 09: HPC for Environmental Computing|
|Time:||Tuesday, June 21, 2016
01:15 pm - 02:15 pm
|Breaks:||12:30 pm - 01:45 pm Lunch|
|Speaker:||Hans-Peter Bunge, LMU|
|Wilco Hazeleger, Netherlands eScience Center|
|Dieter Kranzlmüller, LMU & LRZ|
|Antonio Parodi, CIMA Research Foundation|
|Julio Serje, UNISDR|
|Sanna Sorvari, Finnish Meteorological Institute|
|Abstract:||This BoF brings together environmental computing practitioners to discuss challenges in using high performance computing and big data technologies for their daily work. Ranging from the basic research questions in application domains, via integrating multi-domain, multi-scale models on distributed HPC systems, to demands of inter-governmental collaboration. Environmental computing denotes efforts focusing on producing actionable knowledge by optimally combining different environmental models and data sources over IT infrastructures in general, and HPC systems in particular. This includes various feedback loops in models, applications and systems, and – equally importantly – presenting these relationships in a way that can be readily used both in different decision making processes as well as a foundation for further, cross-disciplinary research activities. Due to the inherent computational demands in this domain (resulting from ever increasing resolution of models as well as the use of probabilistic and ensemble forecasting techniques), environmental computing represents a growing user community for different HPC services. Challenges related to data science, such as data fusion and combination of measurements and simulation results, require comprehensive analysis of characteristics of different combinations of HPC and data management approaches.
The BoF builds on series of previous events and workshops discussing the topic, including:
The workshop is intended for everyone interested in aspects of environmental computing on HPC and Big Data infrastructures. Input from model developers, multi-model specialists and managers of environmental computing data sources are especially welcome, as well as from providers who run applications in the environmental domain.