Strong interactions for precision nuclear physics
Abstract
The scientific research underpinning this SNIC MEDIUM proposal strives to make significant advances in the theoretical description of atomic nuclei. This project is funded within an ERC Starting Grant (2018-2023) "PrecisionNuclei" and a VR project grant (dnr. 2020-05127). In particular, we will use HPC computation to study the prospects of describing atomic nuclei using effective field theories of QCD. Several aspects of this research will require HPC efforts: 1) exploring power-counting schemes in EFT 2) constraining the three-nucleon interaction in chiral EFT using three-nucleon scattering observables 3) computational statistics analyses of strongly interacting systems. All efforts will play a crucial role for predicting the properties of exotic isotopes, strongly interacting matter, neutron star properties.
All employed codes are developed in-house, are well tested, and mainly exploit standard numerical LAPACK routines for optimal parallel (OpenMP-MPI) performance.
The computational challenge encompass:
- efficient fast computation of three-nucleon scattering observables [repeated matrix-vector products and matrix diagonalizations on multiple nodes simultaneously, hybrid OpenMP and MPI codes].
- evaluation of quantum mechanical models for performing many-parameter optimization and inference of coupling constants in nuclear EFT [threaded evaluation of complex likelihoods, parallel evaluation using emulators]
- large-scale matrix diagonalization (to solve the quantum mechanical many-body problem with strong interactions). Sparse MatVec and large VecVec operations.
Our collaboration, that includes researchers in Scandinavia and the US, are currently pioneering such computational nuclear physics efforts.
