Ann Almgren, Lawrence Berkeley National Laboratory
Tue 2:30 – 3:15pm
Block-structured adaptive mesh refinement (AMR) is a powerful tool for improving the computational efficiency and reducing the memory footprint of structured-grid numerical simulations. AMR techniques have been used for over 25 years to solve increasingly complex problems. I will give an overview of what we are doing in Berkeley Lab’s AMR framework, BoxLib, to address the challenges of next-generation multicore architectures and the complexity of multiscale, multiphysics problems, including new ways of thinking about multilevel algorithms and new approaches to data layout and load balancing, in situ and in transit visualization and analytics, and run-time performance modeling and control.