EC3: The impact of the changing nature of computing on climate science 24127
EC3: The impact of the changing nature of computing on climate science
Approved 1
Climate science depends heavily on computer simulations of the real world producing vast amounts of data and executed on a wide variety of platforms. Over the years advances in our ability to simulate the world have piggy backed on Kryder’s and Moore’s Laws, but these laws are dying (and Dennard Scaling, which made it possible for processors to go faster, is over). The obvious industry solution is more parallelism, in supercomputers, cloud computing, storage systems, and everything down into mobiles and edge computing. Alongside all that parallelism is heterogeneity in both storage and compute.

This talk will provide an overview of the interaction of climate modelling with computer technology in the past, and how it might play out in the future given the changing nature of computing. It will become clear that climate modelling is one of the grand computational challenges, stressing our ability to programme for next generation supercomputers and cloud, and stressing our ability to handle data. Current and future solutions to these computer science problems will be discussed, from new ways of data handling, to new maths, the use of artificial intelligence and machine learning, and new ways of programming.

Speaker

Bryan Lawrence, University of Reading
Date & time
Thursday 23 January 2020 7:00pm GMT
End date & time
Thursday 23 January 2020 9:00pm GMT
Venue
University of Reading
Address
Van Emden Theatre
Reading, RG6 6UR
United Kingdom
Category Lecture

Share:

Event Map

Date & time: 23/01/20 19:00:00 GMT