Computer Physics Communications

Quantum computing with qudits, an extension of qubits to multiple levels, offers promising advantages in information representation and computational density. Despite its potential, tools for qudit-based quantum computation remain underdeveloped. This article presents QuForge, a Python-based library that introduces a novel computational framework for simulating quantum circuits with arbitrary qudit dimensions. QuForge distinguishes itself by providing scalable and extensible features, such as a complete set of customizable quantum gates, support for sparse matrix representations, and compatibility with GPU and TPU accelerators to optimize performance. By leveraging differentiable frameworks, QuForge accelerates simulations and facilitates quantum machine learning and algorithm development research. Through empirical demonstrations and benchmarks, we highlight the capabilities of the library to address scalability challenges and enable advances in quantum information science, establishing it as a potential tool for advancing research and applications in high-dimensional quantum computing.

In this paper, we describe a continuous-time quantum walk (CTQW) simulation package for Python 3, covering their theoretical foundations and practical applications. The software provides both unitary and open system evolution of over general graphs, alongside tools for visualization and exploration of several different aspects of the quantum walk. We go over installation, design and performance of the package, concluding with several examples on how QWAK can be used to explore problems such as search, perfect state transfer, among others. Additionally, we demonstrate how CuPy is utilized to leverage GPU acceleration.

The third release of the Quantum Dissipative Dynamics (QDD) package follows the second release [P. M. Dinh, et al., Comp. Phys. Comm. 295 (2024) 108947] in which the focus was mainly made on the optimization of the electronic dynamics, in particular with a CUDA fortran coding to allow the use of a GPU. In this new release, we pursued the exportation of other parts of QDD on CUDA-capable GPUs, as the ionic motion, the coupling of the electrons with a laser field and/or with the ions (via pseudopotentials), and all electronic observables, including the involved photo-electron spectra, energy- and/or angle-resolved. Several specific CUDA optimisations have also been implemented, to improve the performance and the memory usage while keeping the accuracy of the results.

We present STREAmS-2.1, an updated version of the flow solver STREAmS [1], lastly updated in Bernardini et al. Comput. Phys. Commun. 285 (2023) 108644. STREAmS-2.1 merges the features of the curvilinear solver FLEW [2] which is able to simulate three canonical cases, namely the circular arc channel, the curved boundary layer and the airfoil case. Moreover, three new backends are included, i.e., OpenMP (for CPUs), HIP (for AMD GPUs) and OpenMP-offload (tested on Intel GPUs but potentially portable). Finally, in situ visualization layer based on Catalyst2 technology is integrated into the solver to reduce the visualization effort, especially for huge computational grids.

The KM3NeT Collaboration has tackled a common challenge faced by the astroparticle physics community, namely adapting the experiment-specific simulation software to work with the CORSIKA air shower simulation output. The proposed solution is an extension of the open source code gSeaGen, which allows the transport of muons generated by CORSIKA to a detector of any size at an arbitrary depth. The gSeaGen code was not only extended in terms of functionality but also underwent a thorough redesign of the muon propagation routine, resulting in a more accurate and efficient simulation. This paper presents the capabilities of the new gSeaGen code as well as prospects for further developments.

The Optical Bloch Equations (OBEs) are useful for calculating the evolution of the density matrix of an atomic ensemble under the action of some Hamiltonian. A common situation concerns atoms with hyperfine structure that interact with an external magnetic field and laser radiation. When the spectral linewidth of the laser radiation is much larger than the natural linewidth of the transition, the mode spacing is much smaller than the natural linewidth, and the spectral linewidth is much larger than the characteristic evolution time of the density matrix, the OBEs can be reduced to rate equations for Zeeman coherences. We present a toolkit for solving these rate equations based on the QuantumOptics.jl package in the Julia language. Using these tools makes the code much more readable than previous implementations in C/C++, but almost as fast and easier to parallelize. The toolkit includes functions for calculating the steady-state solution of density matrix of alkali metal atoms in the presence of an external magnetic field and exposed to a pump laser beam of arbitrary polarization and propagation direction. Based on this density matrix, the toolkit offers functions to determine the fluorescence intensity of arbitrary polarization and direction as well as the absorption of a weak probe beam, also of arbitrary polarization and propagation direction. It can also produce a plot of the electronic angular momentum distribution of the atom based on the calculated density matrix. The toolkit is available on Github and has been validated by comparing its results to legacy code written in C/C++ and experimental measurements. As a test case, we show how the toolkit can be used to optimize a simple atomic magnetometer.

Goupil is a software library designed for the Monte Carlo transport of low-energy gamma-rays, such as those emitted from radioactive isotopes. The library is distributed as a Python module. It implements a dedicated backward sampling algorithm that is highly effective for geometries where the source size largely exceeds the detector size. When used in conjunction with a conventional Monte Carlo engine (i.e., Geant4), the response of a scintillation detector to gamma-active radio-isotopes scattered over the environment is accurately simulated (to the nearest percent) while achieving events rates of a few kHz (with a ∼2.3 GHz CPU).

A major update of the program FeynGame is introduced. One of its main new functionalities is to visualize Feynman graphs generated by qgraf. The qgraf output can be either pasted into the FeynGame canvas for individual graphs, or the whole qgraf output file can be processed. In addition, a number of new features and improvements have been implemented into FeynGame 3.0 in order to further facilitate the efficient drawing of Feynman diagrams in publication quality. FeynGame is freely available • as jar or MacOS app file from https://web.physik.rwth-aachen.de/user/harlander/software/feyngame • as source code from https://gitlab.com/feyngame/FeynGame

We introduce the program ‘Chapidif’ by describing a study of the properties of aluminum based on simple model dielectric functions. These are generally not available from first principle, and one is forced to describe them in terms of (a sum of) model dielectric functions. The Chapidif program is used to visualize these, check their sum rules and the mathematical relation between the real and imaginary part. In addition, several properties related to the interaction of charged particles (here either protons or electrons) with matter are derived and compared with experiment. By having a single program that can calculate a range of properties, it becomes easy to ensure that the model used is not just able to describe a single observable, but it is transferable, i.e. describes reasonably well a larger range of material properties. A reflection electron energy loss measurement is used as an example of how a comparison of calculated results with experiment can be used to improve the model and thus enhance the quality of the properties derived from the dielectric function.

We present the Materials Learning Algorithms (MALA) package, a scalable machine learning framework designed to accelerate density functional theory (DFT) calculations suitable for large-scale atomistic simulations. Using local descriptors of the atomic environment, MALA models efficiently predict key electronic observables, including local density of states, electronic density, density of states, and total energy. The package integrates data sampling, model training and scalable inference into a unified library, while ensuring compatibility with standard DFT and molecular dynamics codes. We demonstrate MALA's capabilities with examples including boron clusters, aluminum across its solid-liquid phase boundary, and predicting the electronic structure of a stacking fault in a large beryllium slab. Scaling analyses reveal MALA's computational efficiency and identify bottlenecks for future optimization. With its ability to model electronic structures at scales far beyond standard DFT, MALA is well suited for modeling complex material systems, making it a versatile tool for advanced materials research.