Session P14: International Cooperative Efforts for Electronic Structure Methods

The European Theoretical Spectroscopy Facility: an illustration for the power of collective research

As researchers and citizens, we should contribute to facing the grand challenges of our epoch. It is important to work on problems such as climate change or limited ressources. However, maybe the biggest challenge is to find ways to unite our forces and develop models of collaborative problem solving. This is mandatory to deal with complex problems, and it can boost efficiency in any case. Code development is just one example where a constructive and well-organized collaboration can take us much further than individual attempts. On the background of this general idea, we will analyze the impact of the European Theoretical Spectroscopy Facility (ETSF, www.etsf.eu) on the day-to-day research of its members, on the theoretical and computational tools that are produced, and on a wider field of theoretical or experimental research. We will see that much can be learnt from this attempt to consider ideas in competition, with people in collaboration.   

The ABINIT software project

The ABINIT software project aims at providing the total energy, charge density and electronic structure of systems made of electrons and nuclei (molecules and periodic solids) thanks to a ”first-principle” approach. The ground state properties are calculated in the framework of the Density-Functional Theory (DFT). The excited states properties are computed within the Many-Body Perturbation Theory (MBPT). Finally, the response properties are obtained from Density-Functional Perturbation Theory (DFPT). \\ The ABINIT software project was started in 1997 as an open software project, without a ‘‘definite’’ goal, developed using several software engineering techniques to allow international collaboration of many different groups. The first publicly available version of ABINIT was released in December 2000 under the GNU GPL. The software has already been described in various articles~[1-4]. The last stable version of the package (7.10.4) has now a 70 MBytes size, consisting in nearly 1400 files written in F90 (830000 lines) and including documentation, tutorials and more than one thousand tests. The code is developed by an always opened community (around fifty people) and it is used by more than a thousand individuals worldwide. \vspace{5mm}\\ \textbf{References} \begin{enumerate} \item X. Gonze \textit{et al.}, Comp. Mat. Sci. \textbf{25}, 478–492 (2002). \item X. Gonze \textit{et al.}, Z. Kristallogr. \textbf{220}, 558–562 (2005). \item X. Gonze \textit{et al.}, Comput. Phys. Commun. \textbf{180}, 2582-2615 (2009). \item X. Gonze \textit{et al.}, in preparation (2015). \end{enumerate}   

The Road to Interoperable Simulation Software: Examples Using the Qbox Code

The diversity of available simulation software implementing various methods—--from atomistic classical molecular dynamics to quantum many-body perturbation theory—--makes it highly desirable to couple these codes in a seamless fashion. We discuss the approach taken with the Qbox code to couple first-principles molecular dynamics with advanced sampling algorithms and with GW electronic structure calculations. \\{} http://qboxcode.org \\{} http://www.quantum-simulation.org   

The CECAM Electronic Structure Library: community-driven development of software libraries for electronic structure simulations

The CECAM Electronic Structure Library (ESL) is a community-driven effort to segregate shared pieces of software as libraries that could be contributed and used by the community. Besides allowing to share the burden of developing and maintaining complex pieces of software, these can also become a target for re-coding by software engineers as hardware evolves, ensuring that electronic structure codes remain at the forefront of HPC trends. In a series of workshops hosted at the CECAM HQ in Lausanne, the tools and infrastructure for the project were prepared, and the first contributions were included and made available online (http://esl.cecam.org). In this talk I will present the different aspects and aims of the ESL and how these can be useful for the electronic structure community.   

The Yambo code: a comprehensive tool to perform ab-initio simulations of equilibrium and out-of-equilibrium properties

Density functional theory and many-body perturbation theory methods (such as GW and Bethe-Selpether equation) are standard approaches to the equilibrium ground and excited state properties of condensed matter systems, surfaces, molecules and other several kind of materials. At the same time ultra-fast optical spectroscopy is becoming a widely used and powerful tool for the observation of the out-of-equilibrium dynamical processes. In this case the theoretical tools (such as the Baym-Kadanoff equation) are well known but, only recently, have been merged with the ab-Initio approach. And, for this reason, highly parallel and efficient codes are lacking. Nevertheless, the combination of these two areas of research represents, for the ab-initio community, a challenging prespective as it requires the development of advanced theoretical, methodological and numerical tools. Yambo is a popular community software implementing the above methods using plane-waves and pseudo-potentials. Yambo is available to the community as open-source software, and oriented to high-performance computing. The Yambo project aims at making the simulation of these equilibrium and out-of-equilibrium complex processes available to a wide community of users. Indeed the code is used, in practice, in many countries and well beyond the European borders. Yambo is a member of the suite of codes of the MAX European Center of Excellence (Materials design at the exascale) . It is also used by the user facilities of the European Spectroscopy Facility and of the NFFA European Center (nanoscience foundries & fine analysis). In this talk I will discuss some recent numerical and methodological developments that have been implemented in Yambo towards to exploitation of next generation HPC supercomputers. In particular, I will present the hybrid MPI+OpenMP parallelization and the specific case of the response function calculation. I will also discuss the future plans of the Yambo project and its potential use as tool for science dissemination, also in third world countries.