Coupling Approaches for Next-Generation Architectures
New high-performance coupling approaches and capabilities for coupled Earth System Models on next generation computing architectures.
Who We Are
The CANGA project is a joint effort funded by the United States Department of Energy's Office of Science under the Scientific Discovery through Advanced Computing (SciDAC) program. It includes scientists from:
What We Do
Complex Models, Complex Architectures, High Performance.
CANGA is taking a new approach to coupled Earth System Models. New programming models and new algorithms will improve the performance, veracity and accuracy of Earth system simulations.
Tasking Programming Models
High-performance for the Next Generation
Task-based programming models appear to be well suited to next-generation and exascale computing architectures, providing enhanced parallelism, load-balancing, fault tolerance and support of hybrid processing elements. We will be creating new task-based prototypes of coupled Earth System Models to achieve maximum performance and better manage architectural and scientific complexity.
New Tools for Remapping Irregular and Time-Varying Meshes
As we introduce new irregular meshes for the land surface and time varying meshes to capture changing coastlines and sea-level rise, improved remapping capabilities are needed to accurately transfer field data between components in a manner that preserves properties of those fields.
Accurate and Stable Coupling Strategies
Coupled Earth System Models are often pieced together from individual components with a coupling strategy based on simple timescale arguments or maximizing concurrency, with unforeseen consequences. We will be formulating mathematically sound, stable and accurate coupling strategies based on detailed analyses of the coupled mathematical system.