We apply for continuation of large scale NAISS allocation of supercomputer resources to advance thin film physics research at the very international front, for computational modeling of materials and method development. 15 theoretical and as many experimental scientists are active in material science computational research coupled with laboratory testing and verification. We have demonstrated over a number of years that the allocated resources are used in a most efficient and productive manner, resulting in a large number of high-impact scientific publications as well as recognition in the form of grants from VR, SSF, Linköping University, Swedish Government Strategic Research Area, KAW Foundation, Swedish Energy Agency, and more, as well as from interational scientific awards. Motivated by our recent discoveries, the team now proposes to increase research thrust. As our academic output is increasing, we request allocations on a level above what was granted to us in the last round, where we proved to be productive users. Our main software is VASP, which is installed and optimized to scale well up to hundreds of cores, and is supported on all NAISS-centers to which we apply. In the coming allocation period we will concentrate our efforts around these areas: 1. Properties and phase stability of new multifunctional materials. In particular 2D noble metals, like our recently discovered goldene (published in Nature Synthesis), and boron-rich alloys, borides, nitrides, carbides, oxides, high-entropy alloys and compounds, steels, metal/ceramic and molecular-nanolayer-based superlattices will be investigated in terms of phase stability and transformations, fracture resistance, defects, piezoelectric and thermoelectric properties, and spectroscopic response. Theoretical results, supported by machine learning methods (developed by us), will be subject to experimental verification. 2. Method developments for disordered materials, thin film growth, and magnetic materials. We will continue developing our coupled atomistic spin dynamics – ab initio molecular dynamics method and apply it to steels and materials with magnetocaloric prospects. Employment of machine learning potentials will enable large-scale simulations with unprecedented accuracy for disordered materials and thin film growth. 3. Properties of novel superconducting materials. Engineering of superconducting properties in borides and carbides with strains from nanostructuring will be pursued, aiming at increasing the critical temperature. Investigation of radiation hardness and defect annealing processes of unconventional superconductors (cuprates) will be carried out with machine learning potentials and radiation damage models for nuclear fusion applications. The project has implications for societal needs in terms of energy harvesting and production, wear-resistant coatings, magnetic storage media, and neutron detector applications at the ESS in Lund.