Second release of MAX software: report on lighthouse codes' deployment at exascale. Deliverable D1.4 of the HORIZON-EUROHPC-JU-2021-COE-01 project MaX (101093374)

This document presents the second release report of the MAX lighthouse codes in the current phase of the MAX Centre of Excellence. It follows from Deliverable D1.2, which laid the groundwork with the first set of production-ready, portable, and exascaleprepared versions of the core codes, and now documents the transition to innovative, disruptive scientific applications at the exascale frontier.In this release, the focus shifts towards demonstrating how the MAX codes evolve and adapt to real-world scientific grand challenges when deployed on state-of-the-art HPC architectures. Across the different codes – BigDFT, FLEUR, QUANTUM ESPRESSO, SIESTA, and YAMBO – this document highlights:• Substantial performance and scalability improvements, including GPU acceleration and advanced numerical interfaces;• The adoption of modern programming models (OpenACC in QUANTUM ESPRESSO, improved solver abstraction in FLEUR, concurrent CUDA-Fortran, OpenACC, and OpenMP backends in YAMBO);• Integration of new features to support research in fields such as phononics, excitonics, electrochemistry, and nanostructured materials (SIESTA, YAMBO),• The development of non-trivial scientific workflows exploiting hybrid quantum/classical dynamics at scale (BigDFT),• Systematic efforts for deployment and validation across major EuroHPC platforms, ensuring reproducibility, portability, and user support.A consistent emphasis is placed on the algorithmic innovations and software engineering practices needed to achieve exascale performance and robustness. In parallel, each team reflects on open challenges and outlines strategies for future development.This report is structured on a code basis, allowing each team to independently present their achievements and challenges, despite the work is the result of integrated discussions and input from HPC centres at the CoE level. A concluding section provides a synthesis of the overall progress and a forward-looking perspective on the roadmap to exascale-erascientific discovery.