@inproceedings{63890, abstract = {{The computation of highly contracted electron repulsion integrals (ERIs) is essential to achieve quantum accuracy in atomistic simulations based on quantum mechanics. Its growing computational demands make energy efficiency a critical concern. Recent studies demonstrate FPGAs’ superior performance and energy efficiency for computing primitive ERIs, but the computation of highly contracted ERIs introduces significant algorithmic complexity and new design challenges for FPGA acceleration.In this work, we present SORCERI, the first streaming overlay acceleration for highly contracted ERI computations on FPGAs. SORCERI introduces a novel streaming Rys computing unit to calculate roots and weights of Rys polynomials on-chip, and a streaming contraction unit for the contraction of primitive ERIs. This shifts the design bottleneck from limited CPU-FPGA communication bandwidth to available FPGA computation resources. To address practical deployment challenges for a large number of quartet classes, we design three streaming overlays, together with an efficient memory transpose optimization, to cover the 21 most commonly used quartet classes in realistic atomistic simulations. To address the new computation constraints, we use flexible calculation stages with a free-running streaming architecture to achieve high DSP utilization and good timing closure.Experiments demonstrate that SORCERI achieves an average 5.96x, 1.99x, and 1.16x better performance per watt than libint on a 64-core AMD EPYC 7713 CPU, libintx on an Nvidia A40 GPU, and SERI, the prior best-performing FPGA design for primitive ERIs. Furthermore, SORCERI reaches a peak throughput of 44.11 GERIS (109 ERIs per second) that is 1.52x, 1.13x, and 1.93x greater than libint, libintx and SERI, respectively. SORCERI will be released soon at https://github.com/SFU-HiAccel/SORCERI.}}, author = {{Stachura, Philip and Wu, Xin and Plessl, Christian and Fang, Zhenman}}, booktitle = {{Proceedings of the 2026 ACM/SIGDA International Symposium on Field Programmable Gate Arrays (FPGA '26)}}, isbn = {{9798400720796}}, keywords = {{electron repulsion integrals, quantum chemistry, atomistic simulation, overlay architecture, fpga acceleration}}, pages = {{224--234}}, publisher = {{Association for Computing Machinery}}, title = {{{SORCERI: Streaming Overlay Acceleration for Highly Contracted Electron Repulsion Integral Computations in Quantum Chemistry}}}, doi = {{10.1145/3748173.3779198}}, year = {{2026}}, } @inproceedings{65101, abstract = {{Various methods to measure the dynamic behavior of particles require the calculation of autocorrelation functions. For this purpose, fast multi-tau correlators have been developed in dedicated hardware, in software, and on FPGAs. However, for methods such as X-ray Photon Correlation Spectroscopy (XPCS), which requires to calculate the autocorrelation function independently for hundreds of thousands to millions of pixels from high-resolution detectors, current approaches rely on offline processing after data acquisition. Moreover, the internal pipeline state of so many independent correlators is far too large to keep it on-chip. In this work, we propose a design approach on FPGAs, where pipeline contexts are stored in off-chip HBM memory. Each compute unit iteratively loads the state for a single pixel, processes a short time series for this pixel, and afterwards writes back the context in a dataflow pipeline. We have implemented the required compute kernels with Vitis HLS and analyze resulting designs on an Alveo U280 card. The design achieves the expected performance and for the first time provides sufficient throughput for current high-end detectors used in XPCS.}}, author = {{Tareen, Abdul Rehman and Plessl, Christian and Kenter, Tobias}}, booktitle = {{2025 International Conference on Field Programmable Technology (ICFPT)}}, publisher = {{IEEE}}, title = {{{Fast Multi-Tau Correlators on FPGA with Context Switching From and to High- Bandwidth Memory}}}, doi = {{10.1109/icfpt67023.2025.00027}}, year = {{2026}}, } @inproceedings{65267, author = {{Hollenhorst, Viola and Riese, Julia and Kenig, Eugeny Y.}}, location = {{Luzern, Schweiz}}, title = {{{Investigation of Surface Roughness Effects on Flow Patterns and Thermal Performance in Additively Manufactured Channels}}}, year = {{2026}}, } @article{65435, author = {{Devulapally, Deekshith Reddy and Tröster, Thomas}}, journal = {{Discover Mechanical Engineering}}, publisher = {{Springer Nature}}, title = {{{Improving the load-bearing capacity of clinched joints through cavity filling with structural epoxy adhesive.}}}, doi = {{https://doi.org/10.1007/s44245-026-00223-w}}, volume = {{5}}, year = {{2026}}, } @inproceedings{65483, abstract = {{Clinched joints with non-rotationally symmetric geometries exhibit orientation-dependent mechanical behavior that is commonly neglected in structural-scale simulations. Reuleaux triangle shaped clinched joints, in particular, show pronounced in-plane anisotropy depending on their orientation. While such effects have been studied at joint and specimen scale, their relevance at the structural level remains largely unexplored. In this work, the influence of joint orientation on the bending response of a joined structure is investigated using numerical simulations. A simplified joint replacement model based on the *CONSTRAINED_SPR2 point-connector formulation in LS-DYNA is employed, with parameters calibrated from previously obtained experimental force displacement data. A hat shaped profile structure subjected to three-point bending is analyzed in a parametric study considering variations in joint orientation, joint spacing, and profile geometry. The results show that joint orientation has little influence during the initial deformation phase but becomes increasingly significant at larger displacements, where joint behavior governs load transfer. Orientation dependent effects are found to influence the global force displacement response and local load redistribution among joints, with magnitudes comparable to those induced by changes in joint spacing and structural geometry. The findings confirm that joint orientation effects remain relevant at the structural level and should be considered in the design of structures assembled using non-rotationally symmetric clinched joints.}}, author = {{Devulapally, Deekshith Reddy and Tröster, Thomas}}, booktitle = {{Materials Science Forum}}, issn = {{1662-9752}}, pages = {{161--169}}, publisher = {{Trans Tech Publications, Ltd.}}, title = {{{A Numerical Study on the Mutual Influence of Joint Orientation and Component Geometry in Non-Rotationally Symmetric Clinched Joints}}}, doi = {{10.4028/p-0tiihi}}, volume = {{1185}}, year = {{2026}}, } @article{61152, abstract = {{While neural network quantization effectively reduces the cost of matrix multiplications, aggressive quantization can expose non-matrix-multiply operations as significant performance and resource bottlenecks on embedded systems. Addressing such bottlenecks requires a comprehensive approach to tailoring the precision across operations in the inference computation. To this end, we introduce scaled-integer range analysis (SIRA), a static analysis technique employing interval arithmetic to determine the range, scale, and bias for tensors in quantized neural networks. We show how this information can be exploited to reduce the resource footprint of FPGA dataflow neural network accelerators via tailored bitwidth adaptation for accumulators and downstream operations, aggregation of scales and biases, and conversion of consecutive elementwise operations to thresholding operations. We integrate SIRA-driven optimizations into the open-source FINN framework, then evaluate their effectiveness across a range of quantized neural network workloads and compare implementation alternatives for non-matrix-multiply operations. We demonstrate an average reduction of 17\% for LUTs, 66\% for DSPs, and 22\% for accumulator bitwidths with SIRA optimizations, providing detailed benchmark analysis and analytical models to guide the implementation style for non-matrix layers. Finally, we open-source SIRA to facilitate community exploration of its benefits across various applications and hardware platforms.}}, author = {{Umuroglu, Yaman and Berganski, Christoph and Jentzsch, Felix and Danilowicz, Michal and Kryjak, Tomasz and Bezaitis, Charalampos and Sjalander, Magnus and Colbert, Ian and Preusser, Thomas and Petri-Koenig, Jakoba and Blott, Michaela}}, issn = {{1936-7406}}, journal = {{ACM Transactions on Reconfigurable Technology and Systems}}, title = {{{SIRA: Scaled-Integer Range Analysis for Optimizing FPGA Dataflow Neural Network Accelerators}}}, doi = {{10.1145/3807510}}, year = {{2026}}, } @inproceedings{65501, author = {{Stasytis, Lukas and Jentzsch, Felix and Preusser, Thomas and Umuroglu, Yaman and Petri-Koenig, Jakoba and István, Zsolt}}, booktitle = {{2025 International Conference on Field Programmable Technology (ICFPT)}}, publisher = {{IEEE}}, title = {{{Heuristic & Expert-Guided Buffer Sizing for Neural Network Inference Applications on FPGAs}}}, doi = {{10.1109/icfpt67023.2025.00032}}, year = {{2026}}, } @inproceedings{65500, author = {{Jentzsch, Felix and Platzner, Marco}}, booktitle = {{2025 International Conference on Field Programmable Technology (ICFPT)}}, publisher = {{IEEE}}, title = {{{Empirical QoR Estimation Flow for Fast Design Space Exploration of DNN Dataflow Accelerators}}}, doi = {{10.1109/icfpt67023.2025.00044}}, year = {{2026}}, } @article{63744, abstract = {{Orbital angular momentum (OAM) modes are an important resource used in various branches of quantum science and technology due to their unique helical structure and countably infinite basis. Generating light that simultaneously carries high-order orbital angular momenta and exhibits quantum correlations is a challenging task. In this work, we present a theoretical approach to the generation of correlated Schmidt modes carrying OAM via parametric down-conversion (PDC) in cascaded nonlinear systems (nonlinear interferometers) pumped by Laguerre–Gaussian beams. We demonstrate how the number of generated modes and their population can be controlled by varying the pump parameters, the gain of the PDC process, and the distance between the crystals. We investigate the angular displacement measurement uncertainty of these interferometers and demonstrate that it can overcome the classical shot noise limit.}}, author = {{Scharwald, Dennis and Gehse, Lucas and Sharapova, Polina}}, issn = {{2378-0967}}, journal = {{APL Photonics}}, number = {{1}}, publisher = {{AIP Publishing}}, title = {{{Schmidt modes carrying orbital angular momentum generated by cascaded systems pumped with Laguerre–Gaussian beams}}}, doi = {{10.1063/5.0229802}}, volume = {{10}}, year = {{2025}}, } @unpublished{64071, abstract = {{Stimulated by the renewed interest and recent developments in semi-empirical quantum chemical (SQC) methods for noncovalent interactions, we examine the properties of liquid water at ambient conditions by means of molecular dynamics (MD) simulations, both with the conventional NDDO-type (neglect of diatomic differential overlap) methods, e.g. AM1 and PM6, and with DFTB-type (density-functional tight-binding) methods, e.g. DFTB2 and GFN-xTB. Besides the original parameter sets, some specifically reparametrized SQC methods (denoted as AM1-W, PM6-fm, and DFTB2-iBi) targeting various smaller water systems ranging from molecular clusters to bulk are considered as well. The quality of these different SQC methods for describing liquid water properties at ambient conditions are assessed by comparison to well-established experimental data and also to BLYP-D3 density functional theory-based ab initio MD simulations. Our analyses reveal that static and dynamics properties of bulk water are poorly described by all considered SQC methods with the original parameters, regardless of the underlying theoretical models, with most of the methods suffering from too weak hydrogen bonds and hence predicting a far too fluid water with highly distorted hydrogen bond kinetics. On the other hand, the reparametrized force-matchcd PM6-fm method is shown to be able to quantitatively reproduce the static and dynamic features of liquid water, and thus can be used as a computationally efficient alternative to electronic structure-based MD simulations for liquid water that requires extended length and time scales. DFTB2-iBi predicts a slightly overstructured water with reduced fluidity, whereas AM1-W gives an amorphous ice-like structure for water at ambient conditions.}}, author = {{Wu, Xin and Elgabarty, Hossam and Alizadeh, Vahideh and Henao Aristizabal, Andres and Zysk, Frederik and Plessl, Christian and Ehlert, Sebastian and Hutter, Jürg and Kühne, Thomas D.}}, title = {{{Benchmarking semi-empirical quantum chemical methods on liquid water}}}, year = {{2025}}, } @article{63745, abstract = {{Multimode squeezed light is an increasingly popular tool in photonic quantum technologies, including sensing, imaging, and computation. Meanwhile, the existing methods of its characterization are technically complicated, which reduces the level of squeezing, and mostly deal with a single mode at a time. Here, for the first time, to the best of our knowledge, we employ optical parametric amplification to characterize multiple squeezing eigenmodes simultaneously. We retrieve the shapes and squeezing degrees of all modes at once through direct detection followed by modal decomposition. This method is tolerant to inefficient detection and does not require a local oscillator. For a spectrally and spatially multimode squeezed vacuum, we characterize eight strongest spatial modes, obtaining squeezing and anti-squeezing values of up to −5.2 ± 0.2 dB and 8.6 ± 0.3 dB, respectively, despite the 50% detection loss. This work, being the first exploration of an optical parametric amplifier’s multimode capability for squeezing detection, paves the way for the real-time detection of multimode squeezing.}}, author = {{Barakat, Ismail and Kalash, Mahmoud and Scharwald, Dennis and Sharapova, Polina and Lindlein, Norbert and Chekhova, Maria}}, issn = {{2837-6714}}, journal = {{Optica Quantum}}, number = {{1}}, publisher = {{Optica Publishing Group}}, title = {{{Simultaneous measurement of multimode squeezing through multimode phase-sensitive amplification}}}, doi = {{10.1364/opticaq.524682}}, volume = {{3}}, year = {{2025}}, } @inproceedings{59441, abstract = {{Abstract. Accurate Finite Element Modeling (FEM) of joints is essential in the design of complex mechanical systems such as automotive body-in-white (BIW) structures, as it plays a critical role in evaluating their performance. Although well-established techniques exist for modeling rotationally symmetric joints, there remains a significant gap in effectively modeling non-rotationally symmetric joints. These joints are particularly relevant in the automotive BIW, where they can better accommodate anisotropic loading conditions. In this study, strategies for modeling non-rotationally symmetric joints were explored using finite element simulations in LS-DYNA. The findings demonstrate that discrete beam elements can capture the anisotropic characteristics of such joints. Two models were tested: a single-beam model for stiffness periodicity every 90°, and a three-beam model for stiffness periodicity every 120°. Force responses, stress distribution, and sheet bending behaviors were analyzed, confirming that discrete beam elements can accurately represent direction-dependent stiffness. These results establish a foundation for developing advanced joint modeling strategies in complex mechanical systems.}}, author = {{Devulapally, Deekshith Reddy and Tröster, Thomas}}, booktitle = {{Materials Research Proceedings}}, issn = {{2474-395X}}, location = {{Paderborn}}, publisher = {{Materials Research Forum LLC}}, title = {{{Modelling strategies for non-rotationally symmetric joints}}}, doi = {{10.21741/9781644903551-21}}, volume = {{52}}, year = {{2025}}, } @article{58759, author = {{Dechert, Christopher and Kenig, Eugeny Y.}}, issn = {{0009-2509}}, journal = {{Chemical Engineering Science}}, publisher = {{Elsevier BV}}, title = {{{Influence of microstructures on liquid spreading on inclined plates: A CFD based study}}}, doi = {{10.1016/j.ces.2025.121317}}, year = {{2025}}, } @inproceedings{58761, author = {{Dechert, Christopher and Riese, Julia and Franke, Patrick}}, location = {{Bochum}}, title = {{{Untersuchung zur Vergleichbarkeit von verschiedenen Methoden zur Bestimmung der effektiven Phasengrenzflächen}}}, year = {{2025}}, } @article{53805, abstract = {{The article introduces a method to learn dynamical systems that are governed by Euler–Lagrange equations from data. The method is based on Gaussian process regression and identifies continuous or discrete Lagrangians and is, therefore, structure preserving by design. A rigorous proof of convergence as the distance between observation data points converges to zero and lower bounds for convergence rates are provided. Next to convergence guarantees, the method allows for quantification of model uncertainty, which can provide a basis of adaptive sampling techniques. We provide efficient uncertainty quantification of any observable that is linear in the Lagrangian, including of Hamiltonian functions (energy) and symplectic structures, which is of interest in the context of system identification. The article overcomes major practical and theoretical difficulties related to the ill-posedness of the identification task of (discrete) Lagrangians through a careful design of geometric regularisation strategies and through an exploit of a relation to convex minimisation problems in reproducing kernel Hilbert spaces.}}, author = {{Offen, Christian}}, journal = {{Mathematics of Computation}}, publisher = {{American Mathematical Society}}, title = {{{Machine learning of continuous and discrete variational ODEs with convergence guarantee and uncertainty quantification}}}, doi = {{10.1090/mcom/4120}}, year = {{2025}}, } @article{60298, abstract = {{In this work, we introduce PHOENIX, a highly optimized explicit open-source solver for two-dimensional nonlinear Schrödinger equations with extensions. The nonlinear Schrödinger equation and its extensions (Gross-Pitaevskii equation) are widely studied to model and analyze complex phenomena in fields such as optics, condensed matter physics, fluid dynamics, and plasma physics. It serves as a powerful tool for understanding nonlinear wave dynamics, soliton formation, and the interplay between nonlinearity, dispersion, and diffraction. By extending the nonlinear Schrödinger equation, various physical effects such as non-Hermiticity, spin-orbit interaction, and quantum optical aspects can be incorporated. PHOENIX is designed to accommodate a wide range of applications by a straightforward extendability without the need for user knowledge of computing architectures or performance optimization. The high performance and power efficiency of PHOENIX are demonstrated on a wide range of entry-class to high-end consumer and high-performance computing GPUs and CPUs. Compared to a more conventional MATLAB implementation, a speedup of up to three orders of magnitude and energy savings of up to 99.8% are achieved. The performance is compared to a performance model showing that PHOENIX performs close to the relevant performance bounds in many situations. The possibilities of PHOENIX are demonstrated with a range of practical examples from the realm of nonlinear (quantum) photonics in planar microresonators with active media including exciton-polariton condensates. Examples range from solutions on very large grids, the use of local optimization algorithms, to Monte Carlo ensemble evolutions with quantum noise enabling the tomography of the system's quantum state.}}, author = {{Wingenbach, Jan and Bauch, David and Ma, Xuekai and Schade, Robert and Plessl, Christian and Schumacher, Stefan}}, issn = {{0010-4655}}, journal = {{Computer Physics Communications}}, publisher = {{Elsevier BV}}, title = {{{PHOENIX – Paderborn highly optimized and energy efficient solver for two-dimensional nonlinear Schrödinger equations with integrated extensions}}}, doi = {{10.1016/j.cpc.2025.109689}}, volume = {{315}}, year = {{2025}}, } @inproceedings{60494, abstract = {{Pillow-plate heat exchangers (PPHEs) represent a viable alternative to conventional shell-and-tube and plate heat exchangers. The waviness of their channels intensifies fluid mixing in the boundary layers and facilitates heat transfer. Applying secondary surface structuring can further enhance the overall thermo-hydraulic performance of PPHEs, thus increasing their competitiveness against conventional heat exchangers. In this work, streamlined secondary structures applied on the PPHE surface were studied numerically to explore their potential in enhancing near-wall fluid mixing. Computational fluid dynamics (CFD) simulations of single-phase turbulent flow in the inner PPHE channel were performed and pressure drop, heat transfer coefficients, and overall thermo-hydraulic efficiency were determined. The simulation results clearly demonstrate a favourable influence of secondary structuring on the heat transfer performance of PPHEs.}}, author = {{Afsahnoudeh, Reza and Riese, Julia and Kenig, Eugeny Y.}}, booktitle = {{Systems and Control Transactions}}, issn = {{2818-4734}}, publisher = {{PSE Press}}, title = {{{Thermo-Hydraulic Performance of Pillow-Plate Heat Exchangers with Streamlined Secondary Structures: A Numerical Analysis}}}, doi = {{10.69997/sct.178665}}, volume = {{4}}, year = {{2025}}, } @article{60565, author = {{Bocchini, Adriana and Gerstmann, Uwe and Schmidt, Wolf Gero}}, issn = {{2469-9950}}, journal = {{Physical Review B}}, number = {{10}}, publisher = {{American Physical Society (APS)}}, title = {{{Microscopic origin of gray tracks in KTiOPO4}}}, doi = {{10.1103/physrevb.111.104103}}, volume = {{111}}, year = {{2025}}, } @article{58642, abstract = {{We present a cost-effective self-assembly method to fabricate low-density dimer NPs in an NPoM architecture, using the M13 phage as a spacer layer. This will enable the development of dynamic plasmonic devices and advanced sensing applications.}}, author = {{Devaraj, Vasanthan and Ruiz Alvarado, Isaac Azahel and Lee, Jong-Min and Oh, Jin-Woo and Gerstmann, Uwe and Schmidt, Wolf Gero and Zentgraf, Thomas}}, issn = {{2055-6756}}, journal = {{Nanoscale Horizons}}, pages = {{537--548}}, publisher = {{Royal Society of Chemistry (RSC)}}, title = {{{Self-assembly of isolated plasmonic dimers with sub-5 nm gaps on a metallic mirror}}}, doi = {{10.1039/d4nh00546e}}, volume = {{10}}, year = {{2025}}, } @article{61246, abstract = {{Abstract The time-dependent one-dimensional nonlinear Schrödinger equation (NLSE) is solved numerically by a hybrid pseudospectral-variational quantum algorithm that connects a pseudospectral step for the Hamiltonian term with a variational step for the nonlinear term. The Hamiltonian term is treated as an integrating factor by forward and backward Fourier transforms, which are here carried out classically. This split allows us to avoid higher-order time integration schemes, to apply a first-order explicit time stepping for the remaining nonlinear NLSE term in a variational algorithm block, and thus to avoid numerical instabilities. We demonstrate that the analytical solution is reproduced with a small root mean square error for a long time interval over which a nonlinear soliton propagates significantly forward in space while keeping its shape. We analyze the accuracy and complexity of the quantum algorithm, the expressibility of the ansatz circuit and compare it with classical approaches. Furthermore, we investigate the influence of algorithm parameters on the accuracy of the results, including the temporal step width and the depth of the quantum circuit.}}, author = {{Köcher, Nikolas and Rose, Hendrik and Bharadwaj, Sachin S. and Schumacher, Jörg and Schumacher, Stefan}}, issn = {{2045-2322}}, journal = {{Scientific Reports}}, number = {{1}}, publisher = {{Springer Science and Business Media LLC}}, title = {{{Numerical solution of nonlinear Schrödinger equation by a hybrid pseudospectral-variational quantum algorithm}}}, doi = {{10.1038/s41598-025-05660-3}}, volume = {{15}}, year = {{2025}}, }