Searcharxiv⌕ Search

arXiv subjects

Andreas Herten

Publications and source records attributed to Andreas Herten.

At least 19 recordsLinked to original sources

Performance Portability Explained: Insights from accelerating ICON microphysics

Scientific computing is under increasing pressure to adapt to HPC hardware that is shaped by AI workloads. In these modern ecosystems, extracting reasonable performance without investing major engineering effort is a central issue for domain scientists. Correctness constraints and intricate reference implementations represent further impediments to development. Instead of using a double-precision implementation as a reference, considering the underlying scientific problem directly opens more opportunities for performance improvements when updating and rewriting applications. We take seven different portability frameworks---covering pragma-based, DSL, and data-centric approaches---and evaluate them on a microphysics kernel of the ICON climate model. Our performance benchmarks show that each different HPC system benefits from a different combination of optimizations, so we develop the measure of framework expressivity to gauge how easily each framework can customize portable implementations in this way.

cs.DC↗

A GPU-Accelerated Blocked Adaptive Randomized Range Finder Based on an Implicit Householder QR Decomposition

Low-rank methods can reduce the memory and computational requirements of deep neural network training in approaches such as GaLore. Randomized range finders offer an attractive alternative to singular value decompositions, particularly when the required rank is determined adaptively from a prescribed approximation tolerance. We introduce a blocked adaptive randomized range finder based on a Householder QR decomposition and an optimized hybrid CPU--GPU implementation. Numerical experiments show that it preserves orthogonality and approximation accuracy in regimes where block Gram--Schmidt without reorthogonalization becomes unstable. The blocked formulation exposes matrix--matrix operations and enables overlap of CPU panel factorization with GPU updates. On an NVIDIA GH200, the overlapped implementation reduces the runtime for the largest tested matrix from 9.91 s on the CPU to 0.407 s. The method provides a stable and efficient building block for low-rank approximation on heterogeneous systems with applications in computational science and engineering.

math.NA↗

CALIPER: Calibrating HPC/AI Network Simulators

Network simulation plays an important role in HPC/AI system research by enabling exploration of network architectures, communication strategies, and system configurations at scales that may be impractical to study experimentally. Existing approaches to network simulator calibration and validation lack a principled methodology that combines parameter alignment, constrained inference, and independent validation for collective-centric HPC/AI networks. We present CALIPER, a methodology for systematic calibration and validation of HPC/AI network simulators. CALIPER distinguishes directly observable system parameters from auxiliary parameters which are configured via calibration and constrains the latter to feasible domains consistent with the target system. We evaluate CALIPER's hierarchical, scale-aware calibration and its effects on parameter identifiability and generalization using collective communication benchmarks on a reference system with up to 64-nodes at the Jülich Supercomputing Centre. Validation results show that the calibrated parameter configurations identified by CALIPER predict system performance with less than 5% mean errors across scale-up and scale-out network domains. CALIPER demonstrates that hierarchical, constrained calibration can produce simulator configurations whose predictive accuracy persists beyond the observations used for fitting.

cs.NI↗

4D Parallelism Unlocks Exascale Bayesian Neural Networks for High-Fidelity Atmospheric Modeling

We present BEAST, the first-ever Bayesian Swin Transformer for atmospheric forecasting on 0.25$^\circ$ global resolution able to accurately quantify both aleatoric and epistemic uncertainty. To overcome the associated computational bottlenecks, we devise an orthogonal 4D-parallelization scheme that introduces a unique domain-tensor-parallelism strategy and a novel uncertainty parallel method, enabling us to fully leverage GPU capacity and efficiently scale model training. For a 2.4-billion-parameter model, we achieve a peak performance of 3.96 EFLOP/s on 20,480 NVIDIA GH200 GPUs on the JUPITER supercomputer. We train BEAST as a 700-million-parameter model with 96 random weight samples on 384 nodes on 40 years of data for nearly one million gradient updates. This model achieves predictive skill scores competitive with state-of-the-art probabilistic atmospheric AI models and numerical models, and can predict extreme events with exceptional skill, while generating large ensembles 3 to 4 times faster than the current-best AI model. Our contribution unlocks the potential of high-fidelity uncertainty quantification in atmospheric AI models, heralding a new era for AI-based models in climate and Earth system sciences.

physics.ao-ph↗

Characterizing the Scalability and Performance of Large-Scale AI Training Under Multi-Tenancy

Characterising AI workload performance on modern HPC systems requires understanding both their scalability in isolation and their behaviour under concurrent execution. However, the interplay among parallelisation strategies, network congestion, compute capability, and interconnect technologies remains poorly understood. This work investigates the performance and scalability of AI models up to 2400 GPUs. We quantify the communication overheads and their impact across different interconnects by evaluating scale-up, scale-out, and rack-scale configurations under multiple allocation schemes. Finally, we study how multiple concurrent training jobs interfere with each other by designing a realistic noise model. We design a benchmark suite of AI models to evaluate the performance of five distinct parallelisation strategies across different supercomputing clusters, including Alps, Leonardo, LUMI, JUPITER, NVL72 GB300, and DGX A100. Our work provides a systematic characterization of the scalability and execution efficiency of distributed AI training, while offering key insights into performance behavior under realistic multi-tenant scenarios.

cs.DC↗

Performance Evaluation of Fast Fourier Transforms on Emerging RISC-V Hardware with Vector Extension Support

This manuscript presents a performance evaluation of Fast Fourier Transform (FFT) implementations on emerging processors supporting the RISC-V Vector Extension (RVV 1.0). By introducing juFFTe, a light-weight high-performance library for discrete Fourier transforms, it is demonstrated how effective vectorization of performance-critical FFT kernels can be achieved on RVV-enabled hardware. Comprehensive benchmarks on three RVV 1.0-ready processors, the SiFive X280, the X100 core of the SpacemiT K3 and the C920v2 core of the Sophon SG2044, reveal substantial performance improvements of juFFTe (https://github.com/FZJ-JSC/juFFTe) over the widely used FFTW3 library. Although RVV-enabled platforms show promising results at this stage of development, a comparison with AMD's Zen 5 architecture indicates that RISC-V needs further maturing to reach the performance of established micro-architectures.

cs.MS↗

Great Expectations: Benchmarking the Real-World Performance of RVV 1.0 in HPC

Following the ratification of the RISC-V Vector Extension (RVV 1.0), new commercially available silicon has been adopting the extension. This paper revisits the question of RISC-V viability for High-Performance-Computing (HPC) by benchmarking the latest RVV 1.0-capable hardware (SiFive X280 (Tenstorrent Blackhole), SpacemiT X60 (K1) and X100/A100 (K3), and T-Head C920v2 (Sophon SG2044)). We assess these platforms using standard HPC benchmarks (BLAS, FFTW, HPL, HPCG) and synthetic workloads (STREAM, FMA throughput) and compare them to a state-of-the-art HPC ARM64 chip (NVIDIA Grace). Our findings show that while RVV 1.0 delivers significant performance improvements over scalar execution, hardware-specific implementation challenges remain. We detail these performance characteristics and discuss the remaining hurdles for RISC-V, including RVV, to become a mainstay in the HPC landscape.

cs.DC↗

Novel Iterative Construction Methods for the Blocking Job Shop Scheduling Problem

The Blocking Job-Shop Scheduling Problem (BJSSP) arises in modern and complex manufacturing, production, logistics, and service where no intermediate storage is allowed between consecutive operations. This creates a significant challenge for meta-heuristics due to the low ratio of feasible to explored solutions when solving the problem. To address this problem efficiently, we propose three new beam-search-based heuristics: the Beam Search Iterative Construction Heuristic (BS-ICH), its CPU-parallel extension Parallel Multi-Strategy Beam Search (PMS-BS), and a GPU-accelerated variants G-PMS-BS. BS-ICH constructs feasible schedules by iteratively extending partial solutions, while maintaining a beam of width k to preserve multiple high-quality partial schedules. PMS-BS runs hundreds of parallel BS-ICH instances with machine-biased diversity to expand the search space and escape local optima. G-PMS-BS offloads the beam expansion onto massively parallel GPU hardware using a two-phase kernel architecture that separates lightweight scoring from targeted state reconstruction, enabling scaling to instances with 2,000 operations. A hybrid CPU+GPU mode further exploits idle host cores for concurrent exploration, using load-balancing strategy to minimize synchronization overhead. G-PMS-BS achieves a 44x speedup over the CPU baseline. Experiments on all standard Lawrence and Taillard instances demonstrate that G-PMS-BS establishes new best-known results for 22 Lawrence benchmarks and 77 Taillard instances, with makespan reductions of up to 13% on the largest 100x20 instances.

cs.DC↗

Fourier Neural Operators for Rayleigh-Bénard Convection

We propose an improved Fourier Neural Operator (FNO) for modeling two-dimensional Rayleigh-Bénard convection by predicting time increments instead of full solutions, achieving higher accuracy than a standard FNO baseline. The resulting model is compact (314k parameters, 1.26 MB) and fast (7 ms inference), while maintaining similar accuracy as demonstrated in previous benchmarks. We show that although FNOs generalize to finer meshes, accuracy remains limited by the resolution of the training data.

cs.LG↗

Coupling Complementary Simulations for Combined Performance and Energy Optimization

Polymer simulations are among the most computationally demanding workloads in soft-matter research, often requiring days of execution and high energy consumption to achieve physically meaningful results. In this work, we address these challenges through the coupling and optimization of two complementary simulation frameworks: the Uneyama-Doi Model (UDM) and the SOft coarse-grained Monte Carlo Acceleration (SOMA). UDM efficiently propagates concentration fields at the continuum level, while SOMA resolves chain-scale thermal fluctuations via particle-based Monte Carlo dynamics. Each model was individually optimized for GPU execution using kernel fusion, memory coalescing, asynchronous random-number generation yielding up to 70% (UDM) and 80% (SOMA) performance improvement. The coupling is performed through our proposed coordinator library that orchestrates data exchange and synchronizes time-stepping across multiple GPUs. Further management of coupling workload distribution enabled a 13x overall speedup and 24.5x reduction in total energy usage compared to the SOMA baseline, i. e., 96% energy saving. The proposed hybrid approach maintains the same scientific fidelity while drastically reducing the computational and energy footprint, showcasing the potential of energy-aware, cross-application co-design for sustainable high-performance simulations

cs.DC↗

Universal Quantum Computer Simulation of 50 Qubits on Europe`s First Exascale Supercomputer Harnessing Its Heterogeneous CPU-GPU Architecture

We have developed a new version of the high-performance Jülich universal quantum computer simulator (JUQCS-50) that leverages key features of the GH200 superchips as used in the JUPITER supercomputer, enabling simulations of a 50-qubit universal quantum computer for the first time. JUQCS-50 achieves this through three key innovations: (1) extending usable memory beyond GPU limits via high-bandwidth CPU-GPU interconnects and LPDDR5 memory; (2) adaptive data encoding to reduce memory footprint with acceptable trade-offs in precision and compute effort; and (3) an on-the-fly network traffic optimizer. These advances result in a 16.6-fold speedup over the previous 48-qubit record on the K computer

quant-ph↗

Training LLMs on HPC Systems: Best Practices from the OpenGPT-X Project

The training of large language models (LLMs) requires substantial computational resources, complex software stacks, and carefully designed workflows to achieve scalability and efficiency. This report presents best practices and insights gained from the OpenGPT-X project, a German initiative focused on developing open, multilingual LLMs optimized for European languages. We detail the use of high-performance computing (HPC) systems, primarily JUWELS Booster at JSC, for training Teuken-7B, a 7-billion-parameter transformer model. The report covers system architecture, training infrastructure, software choices, profiling and benchmarking tools, as well as engineering and operational challenges. It includes measured throughput data of various configurations of 3D parallelism during training and the impact of features such as flash attention.

cs.DC↗

Nucleon strange electromagnetic form factors using $N_f=2+1+1$ twisted-mass fermions at the physical point

We present the strange electromagnetic form factors of the nucleon using lattice QCD with $N_f=2+1+1$ twisted mass clover-improved fermions and quark masses tuned to their physical values. Using four ensembles with lattice spacings of $a=0.080$ fm, $0.068$ fm, $0.057$ fm and $0.049$ fm, and similar physical volume, we obtain the continuum limit directly at the physical pion mass. The disconnected strange contributions are computed using high statistics two-point functions combined with stochastic noise mitigation techniques, such as spin-color dilution and hierarchical probing in the estimation of the quark loop. From the momentum dependence of the form factors, we provide the strange electric and magnetic radii, as well as the strange magnetic moment in the continuum limit.

hep-lat↗

Nucleon strange electromagnetic form factors from $N_f=2+1+1$ lattice QCD

We present the nucleon strange electromagnetic form factors using four lattice QCD ensembles with $N_f=2+1+1$ twisted mass clover-improved fermions and quark masses tuned to approximately their physical values. The four ensembles have similar physical volume and lattice spacings of $a=0.080$ fm, $0.068$ fm, $0.057$ fm and $0.049$ fm allowing us to take the continuum limit directly at the physical pion mass point. We compute nucleon three-point correlation functions with high statistics, where the disconnected fermion loops are evaluated stochastically with spin-color dilution and hierarchical probing. We find non-zero values for both electric and magnetic form factors. We extract the strange electric and magnetic radii, as well as the strange magnetic moment in the continuum limit by studying the momentum dependence of the form factors. We also compute the charm electromagnetic form factors within the same setup, which we find to be consistent with zero within the statistical precision of our data.

hep-lat↗

Strangeness of nucleons from $N_f=2+1+1$ lattice QCD

We present the strange electromagnetic form factors of the nucleon using lattice QCD simulations with degenerate light, a strange, and a charm quark in the sea with masses tuned to their physical values. For the first time, the strange electromagnetic form factors are computed at the continuum limit using only ensembles simulated with physical quark masses, eliminating the need for chiral extrapolations and their associated systematic uncertainty. We obtain the momentum transfer dependence of the form factors using the $z$-expansion and provide the strange electric and magnetic radii, as well as the strange magnetic moment. When combining our statistical errors and systematic uncertainties stemming from the momentum transfer dependence fit, our errors are an order of magnitude smaller than those associated with experimental determinations of the strange electromagnetic form factor.

hep-lat↗

Efficient Accelerated Graph Edit Distance Computation on GPU

Graph representation is a powerful abstraction of real-world objects and relations. Computing the Graph Edit Distance (GED) between graphs is critical in domains such as bioinformatics, machine learning, and pattern recognition. GED measures the minimum number of edit operations required to transform one graph into another. However, the high computational complexity of optimal and near-optimal methods limits their applicability to large-scale graphs, making high-performance parallel GED computation essential. To address this, we propose FAST-GED, a fast and scalable open-source framework for GED computation on GPUs. FAST-GED overcomes existing limitations by combining high accuracy with fast execution through GPU-friendly algorithmic design and efficient mapping to GPU hardware, minimizing host-device communication. The implementation is optimized and tested across multiple GPU architectures. We validate FAST-GED on real and synthetic datasets with diverse graph sizes and densities. It achieves speedups of several orders of magnitude over the Python NetworkX library while reaching optimal solutions in most cases. Moreover, it outperforms state-of-the-art approximate methods in both accuracy and scalability. We show that FAST-GED enables broader adoption of GED-based solutions in real-world applications.

cs.DC↗

Scaled Block Vecchia Approximation for High-Dimensional Gaussian Process Emulation on GPUs

Emulating computationally intensive scientific simulations is crucial for enabling uncertainty quantification, optimization, and informed decision-making at scale. Gaussian Processes (GPs) offer a flexible and data-efficient foundation for statistical emulation, but their poor scalability limits applicability to large datasets. We introduce the Scaled Block Vecchia (SBV) algorithm for distributed GPU-based systems. SBV integrates the Scaled Vecchia approach for anisotropic input scaling with the Block Vecchia (BV) method to reduce computational and memory complexity while leveraging GPU acceleration techniques for efficient linear algebra operations. To the best of our knowledge, this is the first distributed implementation of any Vecchia-based GP variant. Our implementation employs MPI for inter-node parallelism and the MAGMA library for GPU-accelerated batched matrix computations. We demonstrate the scalability and efficiency of the proposed algorithm through experiments on synthetic and real-world workloads, including a 50M point simulation from a respiratory disease model. SBV achieves near-linear scalability on up to 512 A100 and GH200 GPUs, handles 2.56B points, and reduces energy use relative to exact GP solvers, establishing SBV as a scalable and energy-efficient framework for emulating large-scale scientific models on GPU-based distributed systems.

cs.DC↗

exaCB: Reproducible Continuous Benchmark Collections at Scale Leveraging an Incremental Approach

The increasing heterogeneity of high-performance computing (HPC) systems and the transition to exascale architectures require systematic and reproducible performance evaluation across diverse workloads. While continuous integration (CI) ensures functional correctness in software engineering, performance and energy efficiency in HPC are typically evaluated outside CI workflows, motivating continuous benchmarking (CB) as a complementary approach. Integrating benchmarking into CI workflows enables reproducible evaluation, early detection of regressions, and continuous validation throughout the software development lifecycle. We present exaCB, a framework for continuous benchmarking developed in the context of the JUPITER exascale system. exaCB enables application teams to integrate benchmarking into their workflows while supporting large-scale, system-wide studies through reusable CI/CD components, established harnesses, and a shared reporting protocol. The framework supports incremental adoption, allowing benchmarks to be onboarded easily and to evolve from basic runnability to more advanced instrumentation and reproducibility. The approach is demonstrated in JUREAP, the early-access program for JUPITER, where exaCB enabled continuous benchmarking of over 70 applications at varying maturity levels, supporting cross-application analysis, performance tracking, and energy-aware studies. These results illustrate the practicality using exaCB for continuous benchmarking for exascale HPC systems across large, diverse collections of scientific applications.

cs.DC↗