Search Results (828)

Carbon capture and storage (CCS) is essential for achieving net-zero emissions, yet amine-based capture systems face significant challenges including high energy penalties (20–30% of power plant output) and operational costs ($50–120/tonne CO${}_2$). This study develops and validates a novel multi-scale Digital Twin (DT) framework integrating Physics-Informed Neural Networks (PINNs) to address these challenges through real-time optimization. The framework combines molecular dynamics, process simulation, computational fluid dynamics, and deep learning to enable real-time predictive control. A key innovation is the sequential training algorithm with domain decomposition, specifically designed to handle the nonlinear transport equations governing CO${}_2$ absorption with enhanced convergence properties.The algorithm achieves prediction errors below 1% for key process variables (R${}^2$> 0.98) when validated against CFD simulations across 500 test cases. Experimental validation against pilot-scale absorber data (12 m packing, 30 wt% MEA) confirms good agreement with measured profiles, including temperature (RMSE = 1.2 K), CO${}_2$ loading (RMSE = 0.015 mol/mol), and capture efficiency (RMSE = 0.6%). The trained surrogate enables computational speedups of up to four orders of magnitude, supporting real-time inference with response times below 100 ms suitable for closed-loop control. Under the conditions studied, the framework demonstrates reboiler duty reductions of 18.5% and operational cost reductions of approximately 31%. Sensitivity analysis identifies liquid-to-gas ratio and MEA concentration as the most influential parameters, with mechanistic explanations linking these to mass transfer enhancement and reaction kinetics. Techno-economic assessment indicates favorable investment metrics, though results depend on site-specific factors. The framework architecture is designed for extensibility to alternative solvent systems, with future work planned for industrial-scale validation and uncertainty quantification through Bayesian approaches. Full article

In nanometer integrated-circuit (IC) manufacturing, advanced technology scaling has intensified the effects of process variations on circuit reliability and performance. Random fluctuations in parameters such as threshold voltage, channel length, and oxide thickness further degrade design margins and increase the likelihood of functional failures. These variations often lead to rare circuit failure events, underscoring the importance of accurate yield estimation and robust design methodologies. Conventional Monte Carlo yield estimation is computationally infeasible as millions of simulations are required to capture failure events with extremely low probability. This paper presents a novel reliability-based circuit design optimization framework that leverages deep transfer learning to improve the efficiency of repeated yield analysis in optimization iterations. Based on pre-trained neural network models from prior design knowledge, we utilize model fine-tuning to accelerate importance sampling (IS) for yield estimation. To improve estimation accuracy, adversarial perturbations are introduced to calibrate uncertainty near the model decision boundary. Moreover, we propose a cascaded batch Bayesian optimization (CBBO) framework that incorporates a smart initialization strategy and a localized penalty mechanism, guiding the search process toward high-yield regions while satisfying nominal performance constraints. Experimental validation on SRAM circuits and amplifiers reveals that CBBO achieves a computational speedup of 2.02×–4.63× over state-of-the-art (SOTA) methods, without compromising accuracy and robustness. Full article

Accurate numerical simulation of debris flows is essential for hazard assessment and early-warning design, yet high-fidelity solvers remain computationally expensive, especially when large ensembles must be explored under epistemic uncertainty in rheology, initial conditions, and topography. At the same time, field observations are typically sparse and heterogeneous, limiting purely data-driven approaches. In this work, we develop a deep-learning Fourier Neural Operator (FNO) as a fast, physics-consistent surrogate for one-dimensional shallow-water debris-flow simulations and demonstrate its application to the Rendinara–Morino system in central Italy. A validated finite-volume solver, equipped with HLLC and Rusanov fluxes, hydrostatic reconstruction, Voellmy-type basal friction, and robust wet–dry treatment, is used to generate a large ensemble of synthetic simulations over longitudinal profiles representative of the study area. The parameter space of bulk density, initial flow thickness, and Voellmy friction coefficients is systematically sampled, and the resulting space–time fields of flow depth and velocity form the training dataset. A two-dimensional FNO in the $(x,t)$ domain is trained to learn the full solution operator, mapping topography, rheological parameters, and initial conditions directly to $h(x,t)$ and $u(x,t)$, thereby acting as a site-specific digital twin of the numerical solver. On a held-out validation set, the surrogate achieves mean relative $L^2$ errors of about 6–$7\%$ for flow depth and 10–$15\%$ for velocity, and it generalizes to an unseen longitudinal profile with comparable accuracy. We further show that targeted reweighting of the training objective significantly improves the prediction of the velocity field without degrading depth accuracy, reducing the velocity error on the unseen profile by more than a factor of two. Finally, the FNO provides speed-ups of approximately $36\times$ with respect to the reference solver at inference time. These results demonstrate that combining physics-based synthetic data with operator-learning architectures enables the construction of accurate, computationally efficient, and site-adapted surrogates for debris-flow hazard analysis in data-scarce environments. Full article

Hyperdimensional Computing (HDC) offers a robust and energy-efficient paradigm for edge intelligence; however, current hardware accelerators are often proprietary, tailored to the target learning task and tightly coupled to specific CPU microarchitectures, limiting portability and adoption. To address this, and democratize the deployment of HDC hardware, we present a general-purpose, plug-and-play accelerator IP that implements the Binary Spatter Code framework as a standalone, host-agnostic module. The design is compliant with the AMBA AXI4 standard and provides an AXI4-Lite control plane and DMA-driven AXI4-Stream datapaths coupled to a banked scratchpad memory. The architecture supports synthesis-time scalability, enabling high-throughput transfers independently of the host processor, while employing microarchitectural optimizations to minimize silicon area. A multi-layer C++ software (GitHub repository commit 3ae3b46) stack running in Linux userspace provides a unified programming model, abstracting low-level hardware interactions and enabling the composition of complex HDC pipelines. Implemented on a Xilinx Zynq XC7Z020 SoC, the accelerator achieves substantial gains over an ARM Cortex-A9 baseline, with primitive-level speedups of up to $431\times$. On end-to-end classification benchmarks, the system delivers average speedups of $68.45\times$ for training and $93.34\times$ for inference. The complete RTL and software stack are released as open-source hardware to support reproducible research and rapid adoption on heterogeneous SoCs. Full article

The Fluid Implicit Particle (FLIP) method is widely adopted in fluid simulation due to its computational efficiency and low dissipation. However, its high computational complexity makes it challenging for traditional CPU architectures to meet real-time requirements. To address this limitation, this work migrates the FLIP method to the GPU using the CUDA framework, achieving a transition from conventional CPU computation to large-scale GPU parallel computing. Furthermore, during particle-to-grid (P2G) mapping, the conventional scattering strategy suffers from significant performance bottlenecks due to frequent atomic operations. To overcome this challenge, we propose a GPU parallelization strategy based on spatial hashing indexing and thread block-level cooperation. This approach effectively avoids atomic contention and significantly enhances parallel efficiency. Through diverse fluid simulation experiments, the proposed GPU-parallelized strategy achieves a nearly 50× speedup ratio compared to the conventional CPU-FLIP method. Additionally, in the P2G stage, our method demonstrates over 30% performance improvement relative to the traditional GPU-based particle-thread scattering strategy, while the overall simulation efficiency gains exceeding 20%. Full article

Achieving reliable and real-time inverse kinematics (IK) for 6-degree-of-freedom (6-DoF) spherical-wrist manipulators remains a significant challenge. Analytical formulations often struggle with complex geometries and modeling errors, and standard numerical solvers (e.g., Levenberg–Marquardt) can stall near singularities or converge slowly. Purely data-driven approaches may require large networks and struggle with extrapolation. In this paper, we propose a low-latency, polynomial-based IK solution for spherical-wrist robots. The method leverages spherical coordinates and low-degree polynomial fits for the first three (positional) joints, coupled with a closed-form analytical solver for the final three (wrist) joints. An iterative partial-derivative refinement adjusts the polynomial-based angle estimates using spherical-coordinate errors, ensuring near-zero final error without requiring a full Jacobian matrix. The method systematically enumerates up to eight valid IK solutions per target pose. Our experiments against Levenberg–Marquardt, damped least-squares, and an fmincon baseline show an approximate 8.1× speedup over fmincon while retaining higher accuracy and multi-branch coverage. Future extensions include enhancing robustness through uncertainty propagation, adapting the approach to non-spherical wrists, and developing criteria-based automatic solution-branch selection. Full article

The increasing core counts and architectural heterogeneity of modern processors make performance optimization insufficient if energy consumption is not simultaneously considered. By providing a novel characterization of how the interaction between hybrid architectures and system software disrupts the traditional correlation between execution speed and energy efficiency, this research study analyzes the performance–energy trade-offs of parallel LU matrix decomposition algorithms implemented in Java, focusing on the Crout and Doolittle variants. This study is conducted on four different platforms, including ARM-based, Hybrid x86, and many-core accelerators. Execution time and speedup are evaluated for varying thread counts, while energy consumption is measured externally to capture whole-system energy usage. Experimental results show that the configuration yielding the maximum speedup does not necessarily minimize energy consumption. While x86 systems showed energy savings exceeding 80% under optimal parallel configurations, the ARM-based platform required distinct thread counts to minimize energy consumption compared with maximizing speed. These findings demonstrate that energy-efficient configurations represent a distinct optimization space that often contradicts traditional performance metrics. In the era of hybrid computing, green software optimization must transition from a simplistic “race-to-sleep” paradigm toward sophisticated, architecture-aware strategies that account for the specific power profiles of heterogeneous cores to achieve truly sustainable high-performance computing. Full article

Assessing the stress–strain state around interacting mining excavations using the finite element method (FEM) is computationally expensive for parametric studies. This study evaluates tabular machine-learning surrogate models for the rapid prediction of full stress–strain fields in fractured rock masses treated as an equivalent continuum. A dataset of 1000 parametric FEM simulations using the elastoplastic generalized Hoek–Brown constitutive model was generated to train Random Forest, LightGBM, CatBoost, and Multilayer Perceptron (MLP) models based on geometric features. The results show that the best models achieve R² scores of 0.96–0.97 for stress components and 0.99 for total displacements. LightGBM and CatBoost provide the optimal balance between accuracy and computational cost, offering speed-ups of 15 to 70 times compared to FEM. While Random Forest yields slightly higher accuracy, it is resource-intensive. Conversely, MLP is the fastest but less accurate. These findings demonstrate that data-driven surrogates can effectively replace repeated FEM simulations, enabling efficient parametric analysis and intelligent design optimization for mine workings. Full article

Achieving optimal daylighting is a cornerstone of sustainable architectural design, impacting energy efficiency and occupant well-being. Fast and accurate prediction during the conceptual phase is crucial but challenging. While physics-based simulations are accurate but slow, existing machine learning methods often rely on restrictive parametric inputs, limiting their application across free-form designs. This study presents a novel, geometry-agnostic framework that uses only building facade unfolding diagrams as input to a Generative Adversarial Network (GAN). Our core innovation is a 2D representation that preserves 3D facade geometry and orientation by “unfolding” it onto the floor plan, eliminating the need for predefined parameters or intermediate features during prediction. A Pix2pixHD model was trained, validated, and tested on a total of 720 paired diagram-simulation images (split 80:10:10). The model achieves high-fidelity visual predictions, with a mean Structural Similarity Index (SSIM) of 0.93 against RADIANCE/Daysim benchmarks. When accounting for the practical time of diagram drafting, the complete workflow offers a speedup of approximately 1.5 to 52 times compared to conventional simulation. This work provides architects with an intuitive, low-threshold tool for rapid daylight performance feedback in early-stage design exploration. Full article

The computational tractability of Integrated Information Theory (IIT) is fundamentally constrained by the exponential cost of identifying the Minimum Information Partition (MIP), which is required to quantify integrated information (Φ). Existing approaches become impractical beyond ~15–20 variables, limiting IIT analyses on realistic neural and complex systems. We introduce GeoMIP, a geometric–topological framework that recasts the MIP search as a graph-based optimization problem on the n-dimensional hypercube graph: discrete system states are modeled as graph vertices, and Hamming distance adjacency defines edges and shortest-path structures. Building on a tensor-decomposed representation of the transition probabilities, GeoMIP constructs a transition-cost (ground cost) structure by dynamic programming over graph neighborhoods and BFS-like exploration by Hamming levels, exploiting hypercube symmetries to reduce redundant evaluations. We validate GeoMIP against PyPhi, ensuring reliability of MIP identification and Φ computation. Across multiple implementations, GeoMIP achieves 165–326× speedups over PyPhi while maintaining 98–100% agreement in partition identification. Heuristic extensions further enable analyses up to ~25 variables, substantially expanding the practical IIT regime. Overall, by leveraging the hypercube’s explicit graph structure (vertices, edges, shortest paths, and automorphisms), GeoMIP turns an intractable combinatorial search into a scalable graph-based procedure for IIT partitioning. Full article

Nowadays various coupling strategies have been developed to combine the strengths of different numerical methods in computational fluid dynamics (CFD), among which the coupled algorithm of the lattice Boltzmann-finite volume method (LBM-FVM) has gained widespread attention. However, research on parallel optimization of LBM-FVM coupled solvers remains limited, mostly focused on independent solvers. In this work, we proposed a flexible framework and optimization schemes to explore the coordinated balance of accuracy-efficiency-hardware adaptability. First, we designed a processor layout strategy to address load imbalance and communication redundancy in the coupled solver. We then developed several parallelization techniques, including LBM restructuring, data reuse, and SIMD optimization for targeted kernels on the most advanced architecture of the Sunway series in China, namely SW26010P heterogeneous many-core processors, which provide hardware architectural advantages well suited for large-scale parallel computational fluid dynamics. Finally, the accuracy of the LBM-FVM coupling simulations was validated through benchmark simulations of 2D/3D lid-driven cavity flow. The results show that our LBM-FVM coupling solver can accurately capture flow characteristics, with vortex structures consistent with experimental data. Additionally, we achieved a 152× speedup for the LBM solver and a 126× speedup for the coupled simulation compared to the standalone FVM simulation on the New Sunway supercomputer system. Our approach marks a milestone in the field of LBM implementations and provides a promising future for coupled algorithms in CFD. Full article

Existing State-of-the-Art (SOTA) methods for situational awareness typically rely on high-bandwidth transmission of raw data or computationally intensive models, which are often impractical for resource-constrained edge devices in unstable communication environments. To address these limitations, this paper introduces a comprehensive framework for Regional Risk Monitoring utilizing unmanned swarm systems. We propose an innovative knowledge distillation approach (SIKD) that leverages both soft label dark knowledge and inter-layer relationships, enabling compressed models to run in real time on edge nodes while maintaining high accuracy. Furthermore, recognition results are fused using Bayesian inference to dynamically update the regional risk level. Experimental results demonstrate the feasibility of the proposed framework. Quantitatively, the proposed SIKD algorithm reduces the model parameters by 52.34% and computational complexity to 44.21% of the original model, achieving a 3× inference speedup on edge CPUs. Furthermore, it outperforms state-of-the-art baseline methods (e.g., DKD and IRG) in terms of convergence speed and classification accuracy, ensuring robust real-time risk monitoring. Full article

In this work, different implementation strategies for a Time-Domain (TD) focusing procedure applied to airborne Synthetic Aperture Radar (SAR) raw data are presented, with the key objective of quantitatively assessing their computing time. In particular, two methodological approaches are proposed: a pixel-wise strategy, which processes each image pixel independently, and a matrix-wise strategy, which handles data blocks collectively. Both strategies are further extended to parallel execution frameworks to exploit multi-threading and multi-node capabilities. The presented analysis is conducted within the context of the airborne SAR infrastructure developed at the Institute for Electromagnetic Sensing of the Environment (IREA) of the National Research Council (CNR) in Naples, Italy. This infrastructure integrates an airborne SAR sensor and a high-performance Information Technology (IT) platform well-tailored to the parallel processing of huge amounts of data. Experimental results indicate an advantage of the pixel-wise strategy over the matrix-wise counterpart in terms of computing time. Furthermore, the adoption of parallel processing techniques yields substantial speedups, highlighting its relevance for time-critical SAR applications. These findings are particularly relevant in operational scenarios that demand a rapid data turnaround, such as near-real-time airborne monitoring in emergency response contexts. Full article

In large-scale Unmanned Aerial Vehicle (UAV) and task environments—particularly those involving obstacles—dimensional explosion remains a significant challenge in Multi-Robot Task Allocation (MRTA). To this end, a novel heuristic MRTA framework based on Topological Graph Construction (TGC) is proposed. First, the physical map is transformed into a pixel map, from which a Generalized Voronoi Graph (GVG) is generated by extracting clearance points, which is then used to construct the topological graph of the obstacle environment. Next, the affiliations of UAVs and tasks within the topological graph are determined to partition different topological regions, and the task value of each topological node is calculated, followed by the first-phase Task Allocation (TA) on these topological nodes. Finally, UAVs within the same topological region with their allocated tasks perform a local second-phase TA and generate the final TA result. The simulation experiments analyze the influence of different pixel resolutions on the performance of the proposed method. Subsequently, robustness experiments under localization noise, path cost noise, and communication delays demonstrate that the total benefit achieved by the proposed method remains relatively stable, while the computational time is moderately affected. Moreover, comparative experiments and statistical analyses were conducted against k-means clustering-based MRTA methods in different UAV, task, and obstacle scale environments. The results show that the proposed method improves computational speed while maintaining solution quality, with the PI-based method achieving speedups of over 60 times and the CBBA-based method over 10 times compared with the baseline method. Full article

As one of the most widely used high-performance kernels, General Matrix Multiplication, or GEMM, plays a pivotal role in diverse application fields. With the growing prevalence of training for Convolutional Neural Networks (CNNs) and Large Language Models (LLMs), the design and implementation of high-efficiency, low-precision GEMM on modern Neural Processing Unit (NPU) platforms are of great significance. In this work, HGEMM for Ascend NPU is presented, which enables collaborative processing of different computation types by Cube units and Vector units. The major contributions of this work are the following: (i) dual-stream pipeline scheduling is implemented, which synchronizes padding operations, matrix–matrix multiplications, and element-wise instructions across hierarchical buffers and compute units; (ii) a suite of tiling strategies and a corresponding strategy selection mechanism are developed, comprehensively accounting for the impacts from M, N, and K directions; and (iii) SplitK as well as ShuffleK methods are raised to address the challenges of memory access efficiency and AI Core utilization. Extensive evaluations demonstrate that our proposed HGEMM achieves an average 3.56× speedup over the CATLASS template-based implementation under identical Ascend NPU configurations, and an average 2.10× speedup relative to the cuBLAS implementation on Nvidia A800 GPUs under general random workloads. It also achieves a maximum computational utilization exceeding 90% under benchmark workloads. Moreover, the proposed HGEMM not only significantly outperforms the CATLASS template-based implementation but also delivers efficiency comparable to the cuBLAS implementation in OPT-based bandwidth-limited LLM inference workloads. Full article