SearcharxivSearch

arXiv subjects

Augustin Blanchet

Publications and source records attributed to Augustin Blanchet.

4 recordsLinked to original sources

Arch\^e, an orbital-free molecular dynamics code for fast production of equations of state

We present Arch\^e, an orbital-free molecular dynamics (OFMD) code designed to produce equations of state (EOS) in the plasma state. Unlike in other OFMD codes, Arch\^e uses a self-consistent field (SCF) approach to compute the electronic density. This allows us to implement two algorithms that accelerate SCF convergence by a factor of up to six. First, the density is initialized using the results from the previous MD timestep to define a one-center profile, which is then applied to the new nuclei positions. Second, the initial and final densities are mixed at each SCF iteration in a proportion that minimizes an approximate free energy. We validate the code by comparing the calculated aluminum EOS to results obtained with the Kohn-Sham density functional theory software Abinit. Achieving agreement in the internal energy requires adding a correction related to the norm-conserving pseudopotential derived from an average-atom model. Performance is compared across CPU and GPU architectures, demonstrating an order-of-magnitude speedup for a single GPU compared to 256 CPUs. Arch\^e exhibits an overall linear computational complexity with respect to the number of atoms, as well as the number of real and reciprocal grid points. Execution time is weakly dependent on density; however, interestingly, it decreases as temperature increases -- in contrast to simulations based on Kohn-Sham orbitals.

physics.plasm-ph

Abinit 2025: New Capabilities for the Predictive Modeling of Solids and Nanomaterials

Abinit is a widely used scientific software package implementing density functional theory and many related functionalities for excited states and response properties. This paper presents the novel features and capabilities, both technical and scientific, which have been implemented over the past 5 years. This evolution occurred in the context of evolving hardware platforms, high-throughput calculation campaigns, and the growing use of machine learning to predict properties based on databases of first principles results. We present new methodologies for ground states with constrained charge, spin or temperature; for density functional perturbation theory extensions to flexoelectricity and polarons; and for excited states in many-body frameworks including GW, dynamical mean field theory, and coupled cluster. Technical advances have extended abinit high-performance execution to graphical processing units and intensive parallelism. Second principles methods build effective models on top of first principles results to scale up in length and time scales. Finally, workflows have been developed in different community frameworks to automate \abinit calculations and enable users to simulate hundreds or thousands of materials in controlled and reproducible conditions.

cond-mat.mtrl-sci

Sommerfeld expansion of electronic entropy in INFERNO-like average atom model

In average atom (AA) model, the entropy provides a route to compute thermal electronic contributions to the equation of state (EOS). The complete EOS comprises in many modelings an additional 0K-isotherm and a thermal ionic part. Even at low temperature, the AA model is believed to be the best practical approach. However, when it comes to determine the thermal electronic EOS at low temperatures, the numerical implementation of AA models faces convergence issues related to the pressure ionization of bound states. At contrast, the Sommerfeld expansion tells us that the variations with temperature of thermodynamic variables should express in simple terms at these low temperatures. This led us to tackle the AA predictions with respect to the Sommerfeld expansion of the electronic entropy. We performed a comprehensive investigation for various chemical elements belonging to $s-$, $p-$, $d-$, and $f-$blocks of the periodic table, at varying densities. This was realized using an Inferno-like model since this approach provides the best theoretical framework to address these issues. Practical prescriptions are provided as functions of the atomic number.

cond-mat.stat-mech

Requirements for very high temperature Kohn-Sham density functional simulations and how to bypass them

In high temperature density functional theory simulations (from tens of eV to keV) the total number of Kohn-Sham orbitals is a critical quantity to get accurate results. To establish the relationship between the number of orbitals and the level of occupation of the highest orbital, we derived a model based on the electron gas properties at finite temperature. This model predicts the total number of orbitals required to reach a given level of occupation and thus a stipulated precision. Levels of occupation as low as 10-4, and below, must be considered to get converged results better than 1%, making high temperature simulations very time consuming beyond a few tens of eV. After assessing the predictions of the model against previous results and ABINIT minimizations, we show how the extended FPMD method of Zhang et al. [PoP 23 042707, 2016] allows to bypass these strong constraints on the number of orbitals at high temperature.

physics.comp-ph