08:25 AM - 09:05 AM (40 min)

Future of the Einstein Toolkit: CarpetX

Lorenzo Ennoggi, Erik Schnetter (Rochester Institute of Technology, Perimeter Institute)

Abstract: CarpetX is a new “driver” for the Einstein Toolkit, using the AMReX exascale adaptive mesh refinement framework. I will describe the role of a driver in general, the new features that CarpetX provides, and how they can be used by physics code. These new features include: code that is simpler to write, various dataflow errors that are now automatically detected, improved discretization methods (e.g. conservation, higher order accuracy), and improved performance (e.g. multi-threading, GPUs).

slides: N/A
recording: watch

Edit on GitHub