Chemists and Physicists learn from each other

MPI participates in new Collaborative Research Center

The DFG is funding a new Collaborative Research Center involving the MPI, Bonn University and the Forschungszentrum Jülich.

Together with Bonn University and the Forschungszentrum Jülich, the Max Planck Institut für Kohlenforschung will be running a new Collaborative Research Center from April 2024: The name "NuMeriQS" stands for Numerical Methods for the Investigation of Dynamics and Structure Formation in Quantum Systems. Scientists from various disciplines will be working on the topic of scientific computing. The project is funded by the DFG (German Research Foundation).

Mathematicians, computer scientists, physicists and chemists are each working on the theory of their respective fields. "The problem is that we often don't know what each other is doing," says Prof. Dr. Frank Neese, Director at the Max Planck Institut für Kohlenforschung and Head of the Department of Molecular Theory and Spectroscopy, describing the situation. Part of the new Collaborative Research Center is intended to address precisely this and remedy the situation - as an intercultural exchange platform, so to speak. "This is specifically about chemists and physicists learning new approaches to solving their complicated equations," says Neese. They often deal with similar questions, but use completely different vocabulary. "The aim, if you like, is for us to speak the same language in future, and to do so fluently."

Computers write their own programs

A second part of the Collaborative Research Center, in which Frank Neese is involved, is concerned with the automatic generation of computer programs on large computer systems. This has to do with the fact that software and hardware have not developed at the same pace when it comes to computing. "In the area of hardware, we have made incredibly big steps forward in recent years," says Frank Neese.

However, this hasn't always been the case with software - the scientific programs in particular are quite outdated. ORCA, for example, the quantum chemistry program developed by Neese himself, is one of the youngest programs, and that dates back to the 1990s. "We run the risk of our programs no longer being compatible with modern hardware."

However, simply rewriting a program like ORCA is not trivial. ORCA contains around 2-3 million lines of code, which - roughly estimated - is the equivalent of a book with around 100,000 pages. "Our goal is to teach computers to write such programs themselves, error-free and highly efficiently," says Neese, summarizing his project.

"NuMeriQS" will initially be funded by the DFG for four years, with the prospect of an extension. In addition to Frank Neese himself, doctoral students and a DFG-funded postdoctoral researcher will also be working on the project at the MPI.