What is it about?

This work introduces a software package, called "modred", that easily and efficiently simplifies large complicated simulations (e.g., climate change simulations). These simulations often require supercomputers. This new software takes data from just a few example large simulations, and finds much simpler approximations (called "reduced order models") that can be run in a few seconds on a laptop. The approximations are sometimes over a million times smaller and faster.

Featured Image

Why is it important?

The smaller and faster approximations ("reduced order models") make it much easier to iterate on different ways to understand and influence the original simulation and system. In the example of climate change simulations, the impact of different types of human actions can be quickly approximated without needing supercomputers.

Perspectives

This software is useful for many different types of problems because it is user friendly, well documented, written in Python, can handle very large data (parallelized), and works with *any* data format with only a little bit of additional code written by the user. Many research groups use it today, and the authors maintain it.

Brandt Belson
Princeton University

Read the Original

This page is a summary of: Algorithm 945, ACM Transactions on Mathematical Software, June 2014, ACM (Association for Computing Machinery),
DOI: 10.1145/2616912.
You can read the full text:

Read

Contributors

The following have contributed to this page