NCAR's Future Post Processing & Analysis Demands of Weather and Climate Science

Register here.   -  NOTE: For folks unable to attend in person register and we will email you a livestream link 2 hours prior to event.

Wednesday, November 9, 2016 - 6:00 PM to 8:30 PM

ATLAS Building - University of Colorado Boulder - 1125 18th St., Boulder, CO

Location: ATLAS - Bldg 223, Boulder, CO - Room 100 - Map: http://goo.gl/maps/XTJ9v

Agenda:

6:00 - 6:25 Schmooze - Food shall be served in Lobby

6:25 - 6:30 Announcements

6:30 - 7:30 NCAR's Future Post Processing and Data Analysis Demands of Weather and Climate Science by Dr. Richard Loft

7:30 - 8:30 Network at The Sink - See: http://thesink.com

NCAR's Future Post Processing and Data Analysis Demands of Weather and Climate Science - Abstract

Climate and weather models have long been considered a "grand challenge" computational problems. In recent years they have also become increasingly recognized as data-intensive challenges, requiring innovations on the analysis side, and a more holistic view of the performance of the end-to-end workflow. These trends have implications for both the software and hardware architecture of NCAR's future modeling enterprise. NCAR's Computational and Information Systems Laboratory (CISL) is currently engaged in a number of projects designed to move climate and weather research for future highly parallel systems with novel memory and data storage architectures. These initiatives include:

Preparing analysis workflows for "Big Data" campaigns such as the Climate Model Inter-comparison Project (CMIP) - part of the Inter-governmental Panel on Climate Change (IPCC) Assessment. These efforts include through workflow parallelism, automation, and I/O optimization.

The CMIP Analysis Platform: an NSF-funded initiative to facilitate the inter-comparison of models results of future climate change scenarios.

Project Capstone: Lowering the barriers to broader use of climate and weather analytics through the use of Docker Containers, Web-enabled, RESTful microservices, and problem solving environments such as Jupyter Notebooks.

Project Zeta: Looking farther out data compression capabilities and the use of novel data-centric architectures will come into play. A vision for a next Generation Climate Computer (Project Zeta) will be presented that pulls many of these elements together into a data-intensive supercomputing system designed to accelerate scientific research.

Dr. Richard Loft - Bio

Dr. Richard Loft has worked in high performance computing since joining Thinking Machine Corporation in 1989, and has worked at the National Center for Atmospheric Research (NCAR) since 1994. At SC2001 he was on a team that received a Gordon Bell prize honorable mention for developing a scalable atmospheric dynamical core called the High Order Method Modeling Environment (HOMME), which was recently integrated as part of the widely used Community Earth System Model. In 2005, Rich was NCAR PI on an NSF project to deploy and evaluate ultra-scalable models on an IBM Blue Gene/L system. Dr. Loft is currently Director of the Technology Development Division (TDD) in the Computational and Information Systems Laboratory (CISL) at NCAR. In this capacity he oversees CISL’s R&D efforts, in areas such as applied computer science, visualization and enabling technologies, and earth system modeling infrastructure. Recognizing the need to engage the next generation in high performance computing, in 2007 Dr. Loft 
created the Summer Internships in Parallel Computational Science program (http://www2.cisl.ucar.edu/siparcs) at NCAR.