arXiv subjects

Rob V. van Nieuwpoort

Publications and source records attributed to Rob V. van Nieuwpoort.

15 recordsLinked to original sources

arXiv2026

Enhancing SLMs for Sustainable Code Optimization in Radio-Astronomy

Recent Large Language Models (LLMs) can produce and optimize complex code. We investigate the use of LLMs to generate and optimize code for large-scale sciences, focusing on radio astronomy and sustainability. The LOFAR telescope is currently being upgraded, significantly increasing the sky area observed, while simultaneously processing more data faster. However, this is expected to increase the computational requirements 40-fold. This upgrade thus critically depends on rigorous performance optimization of existing software and widespread adoption of accelerators. The code base is very large, making this a daunting task. We therefore investigate and demonstrate an AI-driven approach meant to assist developers in evaluating and optimizing their code, including porting to hardware accelerators. The LOFAR community is committed to sustainable solutions, and needs to achieve these improvements without increasing the energy budget. We thus need to optimize existing codes or port them to accelerators, while making sure that the optimization process itself is also energy efficient. This poses a challenge, since LLMs are energy-intensive. We therefore propose to use Small Language Models (SLMs) instead to limit environmental impact. In this paper, we show how to enhance SLMs through the use of agentic AI. We extend the SLMs in two ways to improve code generation quality and performance: first with a multi-sampling generation strategy and second with incorporating compiler feedback. We demonstrate that multi-sampling SLMs can match or surpass larger single-generation models with fewer computational resources and that feeding compiler output back into the SLMs leads to consistent improvements across all tested models. Our approach is generic, and can also use Retrieval Augmented Generation (RAG) as well as static and dynamic analysis tools in the code generation pipeline.

cs.SE↗

arXiv2026

Constraint-aware Optimization in Auto-Tuning

Automatic performance tuning, or auto-tuning, is a key technique in high-performance computing, enabling applications to adapt to complex and evolving hardware architectures. A central challenge is the need to optimize over large discrete, constrained parameter spaces, where many candidate configurations are invalid due to hardware or software correctness constraints. Traditional evolutionary algorithms, such as Differential Evolution, Particle Swarm Optimization, and Genetic Algorithms, are not inherently constraint-aware and thus often waste computational resources evaluating invalid solutions. In this work, we present and evaluate constraint-aware variants of four evolutionary algorithms for auto-tuning. Through extensive experiments on a representative benchmark suite, we show that constraint-aware optimization leads to faster convergence and improved performance over unconstrained methods. Furthermore, we demonstrate that our methods outperform the pyATF methods, a state-of-the-art framework for constraint-based auto-tuning. Our results demonstrate that incorporating constraint-awareness into the optimization process significantly enhances their applicability and effectiveness in real-world auto-tuning problems. Constraint-awareness improved algorithm efficiency by ~39 on average, correlated with search space sparsity. The algorithms developed in this study are publicly available as open-source contributions to the Kernel Tuner framework, facilitating future research and benefitting users.

cs.NE↗

arXiv2025

Tuning the Tuner: Introducing Hyperparameter Optimization for Auto-Tuning

Automatic performance tuning (auto-tuning) is widely used to optimize performance-critical applications across many scientific domains by finding the best program variant among many choices. Efficient optimization algorithms are crucial for navigating the vast and complex search spaces in auto-tuning. As is well known in the context of machine learning and similar fields, hyperparameters critically shape optimization algorithm efficiency. Yet for auto-tuning frameworks, these hyperparameters are almost never tuned, and their potential performance impact has not been studied. We present a novel method for general hyperparameter tuning of optimization algorithms for auto-tuning, thus "tuning the tuner". In particular, we propose a robust statistical method for evaluating hyperparameter performance across search spaces, publish a FAIR data set and software for reproducibility, and present a simulation mode that replays previously recorded tuning data, lowering the costs of hyperparameter tuning by two orders of magnitude. We show that even limited hyperparameter tuning can improve auto-tuner performance by 94.8% on average, and establish that the hyperparameters themselves can be optimized efficiently with meta-strategies (with an average improvement of 204.7%), demonstrating the often overlooked hyperparameter tuning as a powerful technique for advancing auto-tuning research and practice.

cs.LG↗

arXiv2025

Efficient Construction of Large Search Spaces for Auto-Tuning

Automatic performance tuning, or auto-tuning, accelerates high-performance codes by exploring vast spaces of code variants. However, due to the large number of possible combinations and complex constraints, constructing these search spaces can be a major bottleneck. Real-world applications have been encountered where the search space construction takes minutes to hours or even days. Current state-of-the-art techniques for search space construction, such as chain-of-trees, lack a formal foundation and only perform adequately on a specific subset of search spaces. We show that search space construction for constraint-based auto-tuning can be reformulated as a Constraint Satisfaction Problem (CSP). Building on this insight with a CSP solver, we develop a runtime parser that translates user-defined constraint functions into solver-optimal expressions, optimize the solver to exploit common structures in auto-tuning constraints, and integrate these and other advances in open-source tools. These contributions substantially improve performance and accessibility while preserving flexibility. We evaluate our approach using a diverse set of benchmarks, demonstrating that our optimized solver reduces construction time by four orders of magnitude versus brute-force enumeration, three orders of magnitude versus an unoptimized CSP solver, and one to two orders of magnitude versus leading auto-tuning frameworks built on chain-of-trees. We thus eliminate a critical scalability barrier for auto-tuning and provide a drop-in solution that enables the exploration of previously unattainable problem scales in auto-tuning and related domains.

cs.DC↗

arXiv2024

Multi-Strided Access Patterns to Boost Hardware Prefetching

Important memory-bound kernels, such as linear algebra, convolutions, and stencils, rely on SIMD instructions as well as optimizations targeting improved vectorized data traversal and data re-use to attain satisfactory performance. On on temporary CPU architectures, the hardware prefetcher is of key importance for efficient utilization of the memory hierarchy. In this paper, we demonstrate that transforming a memory access pattern consisting of a single stride to one that concurrently accesses multiple strides, can boost the utilization of the hardware prefetcher, and in turn improves the performance of memory-bound kernels significantly. Using a set of micro-benchmarks, we establish that accessing memory in a multi-strided manner enables more cache lines to be concurrently brought into the cache, resulting in improved cache hit ratios and higher effective memory bandwidth without the introduction of costly software prefetch instructions. Subsequently, we show that multi-strided variants of a collection of six memory-bound dense compute kernels outperform state-of-the-art counterparts on three different micro-architectures. More specifically, for kernels among which Matrix Vector Multiplication, Convolution Stencil and kernels from PolyBench, we achieve significant speedups of up to 12.55x over Polly, 2.99x over MKL, 1.98x over OpenBLAS, 1.08x over Halide and 1.87x over OpenCV. The code transformation to take advantage of multi-strided memory access is a natural extension of the loop unroll and loop interchange techniques, allowing this method to be incorporated into compiler pipelines in the future.

cs.PF↗

arXiv2024

SuperCode: Sustainability PER AI-driven CO-DEsign

Currently, data-intensive scientific applications require vast amounts of compute resources to deliver world-leading science. The climate emergency has made it clear that unlimited use of resources (e.g., energy) for scientific discovery is no longer acceptable. Future computing hardware promises to be much more energy efficient, but without better optimized software this cannot reach its full potential. In this vision paper, we propose a generic AI-driven co-design methodology, using specialized Large Language Models (like ChatGPT), to effectively generate efficient code for emerging computing hardware. We describe how we will validate our methodology with two radio astronomy applications, with sustainability as the key performance indicator. This paper is a modified version of our accepted SuperCode project proposal. We present it here in this form to introduce the vision behind this project and to disseminate the work in the spirit of Open Science and transparency. An additional aim is to collect feedback, invite potential collaboration partners and use-cases to join the project.

astro-ph.IM↗

arXiv2022

Learning to detect RFI in radio astronomy without seeing it

Radio Frequency Interference (RFI) corrupts astronomical measurements, thus affecting the performance of radio telescopes. To address this problem, supervised segmentation models have been proposed as candidate solutions to RFI detection. However, the unavailability of large labelled datasets, due to the prohibitive cost of annotating, makes these solutions unusable. To solve these shortcomings, we focus on the inverse problem; training models on only uncontaminated emissions thereby learning to discriminate RFI from all known astronomical signals and system noise. We use Nearest-Latent-Neighbours (NLN) - an algorithm that utilises both the reconstructions and latent distances to the nearest-neighbours in the latent space of generative autoencoding models for novelty detection. The uncontaminated regions are selected using weak-labels in the form of RFI flags (generated by classical RFI flagging methods) available from most radio astronomical data archives at no additional cost. We evaluate performance on two independent datasets, one simulated from the HERA telescope and another consisting of real observations from LOFAR telescope. Additionally, we provide a small expert-labelled LOFAR dataset (i.e., strong labels) for evaluation of our and other methods. Performance is measured using AUROC, AUPRC and the maximum F1-score for a fixed threshold. For the simulated data we outperform the current state-of-the-art by approximately 1% in AUROC and 3% in AUPRC for the HERA dataset. Furthermore, our algorithm offers both a 4% increase in AUROC and AUPRC at a cost of a degradation in F1-score performance for the LOFAR dataset, without any manual labelling.

astro-ph.IM↗

arXiv2022

Improving Novelty Detection using the Reconstructions of Nearest Neighbours

We show that using nearest neighbours in the latent space of autoencoders (AE) significantly improves performance of semi-supervised novelty detection in both single and multi-class contexts. Autoencoding methods detect novelty by learning to differentiate between the non-novel training class(es) and all other unseen classes. Our method harnesses a combination of the reconstructions of the nearest neighbours and the latent-neighbour distances of a given input's latent representation. We demonstrate that our nearest-latent-neighbours (NLN) algorithm is memory and time efficient, does not require significant data augmentation, nor is reliant on pre-trained networks. Furthermore, we show that the NLN-algorithm is easily applicable to multiple datasets without modification. Additionally, the proposed algorithm is agnostic to autoencoder architecture and reconstruction error method. We validate our method across several standard datasets for a variety of different autoencoding architectures such as vanilla, adversarial and variational autoencoders using either reconstruction, residual or feature consistent losses. The results show that the NLN algorithm grants up to a 17% increase in Area Under the Receiver Operating Characteristics (AUROC) curve performance for the multi-class case and 8% for single-class novelty detection.

cs.LG↗

arXiv2020

Deep Learning Assisted Data Inspection for Radio Astronomy

Modern radio telescopes combine thousands of receivers, long-distance networks, large-scale compute hardware, and intricate software. Due to this complexity, failures occur relatively frequently. In this work we propose novel use of unsupervised deep learning to diagnose system health for modern radio telescopes. The model is a convolutional Variational Autoencoder (VAE) that enables the projection of the high dimensional time-frequency data to a low-dimensional prescriptive space. Using this projection, telescope operators are able to visually inspect failures thereby maintaining system health. We have trained and evaluated the performance of the VAE quantitatively in controlled experiments on simulated data from HERA. Moreover, we present a qualitative assessment of the the model trained and tested on real LOFAR data. Through the use of a naive SVM classifier on the projected synthesised data, we show that there is a trade-off between the dimensionality of the projection and the number of compounded features in a given spectrogram. The VAE and SVM combination scores between 65% and 90% accuracy depending on the number of features in a given input. Finally, we show the prototype system-health-diagnostic web framework that integrates the evaluated model. The system is currently undergoing testing at the ASTRON observatory.

astro-ph.IM↗

arXiv2020

Real-Time RFI Mitigation for the Apertif Radio Transient System

Current and upcoming radio telescopes are being designed with increasing sensitivity to detect new and mysterious radio sources of astrophysical origin. While this increased sensitivity improves the likelihood of discoveries, it also makes these instruments more susceptible to the deleterious effects of Radio Frequency Interference (RFI). The challenge posed by RFI is exacerbated by the high data-rates achieved by modern radio telescopes, which require real-time processing to keep up with the data. Furthermore, the high data-rates do not allow for permanent storage of observations at high resolution. Offline RFI mitigation is therefore not possible anymore. The real-time requirement makes RFI mitigation even more challenging because, on one side, the techniques used for mitigation need to be fast and simple, and on the other side they also need to be robust enough to cope with just a partial view of the data. The Apertif Radio Transient System (ARTS) is the real-time, time-domain, transient detection instrument of the Westerbork Synthesis Radio Telescope (WSRT), processing 73 Gb of data per second. Even with a deep learning classifier, the ARTS pipeline requires state-of-the-art real-time RFI mitigation to reduce the number of false-positive detections. Our solution to this challenge is RFIm, a high-performance, open-source, tuned, and extensible RFI mitigation library. The goal of this library is to provide users with RFI mitigation routines that are designed to run in real-time on many-core accelerators, such as Graphics Processing Units, and that can be highly-tuned to achieve code and performance portability to different hardware platforms and scientific use-cases. Results on the ARTS show that we can achieve real-time RFI mitigation, with a minimal impact on the total execution time of the search pipeline, and considerably reduce the number of false-positives.

astro-ph.IM↗

arXiv2019

On optimising cost and value in compute systems for radio astronomy

Large-scale science instruments, such as the distributed radio telescope LOFAR, show that we are in an era of data-intensive scientific discovery. Such instruments rely critically on significant computing resources, both hardware and software, to do science. Considering limited science budgets, and the small fraction of these that can be dedicated to compute hardware and software, there is a strong and obvious desire for low-cost computing. However, optimising for cost is only part of the equation; the value potential over the lifetime of the solution should also be taken into account. Using a tangible example, compute hardware, we introduce a conceptual model to approximate the lifetime relative science value of such a system. While the introduced model is not intended to result in a numeric value for merit, it does enumerate some components that define this metric. The intent of this paper is to show how compute system related design and procurement decisions in data-intensive science projects should be weighed and valued. By using both total cost and science value as a driver, the science output per invested Euro is maximised. With a number of case studies, focused on computing applications in radio astronomy past, present and future, we show that the hardware-based analysis can be, and has been, applied more broadly.

astro-ph.IM↗

arXiv2017

Correlating Radio Astronomy Signals with Many-Core Hardware

A recent development in radio astronomy is to replace traditional dishes with many small antennas. The signals are combined to form one large, virtual telescope. The enormous data streams are cross-correlated to filter out noise. This is especially challenging, since the computational demands grow quadratically with the number of data streams. Moreover, the correlator is not only computationally intensive, but also very I/O intensive. The LOFAR telescope, for instance, will produce over 100 terabytes per day. The future SKA telescope will even require in the order of exaflops, and petabits/s of I/O. A recent trend is to correlate in software instead of dedicated hardware, to increase flexibility and to reduce development efforts. We evaluate the correlator algorithm on multi-core CPUs and many-core architectures, such as NVIDIA and ATI GPUs, and the Cell/B.E. The correlator is a streaming, real-time application, and is much more I/O intensive than applications that are typically implemented on many-core hardware today. We compare with the LOFAR production correlator on an IBM Blue Gene/P supercomputer. We investigate performance, power efficiency, and programmability. We identify several important architectural problems which cause architectures to perform suboptimally. Our findings are applicable to data-intensive applications in general. The processing power and memory bandwidth of current GPUs are highly imbalanced for correlation purposes. While the production correlator on the Blue Gene/P achieves a superb 96% of the theoretical peak performance, this is only 16% on ATI GPUs, and 32% on NVIDIA GPUs. The Cell/B.E. processor, in contrast, achieves an excellent 92%. We found that the Cell/B.E. and NVIDIA GPUs are the most energy-efficient solutions, they run the correlator at least 4 times more energy efficiently than the Blue Gene/P.

astro-ph.IM↗

arXiv2017

Towards exascale real-time RFI mitigation

We describe the design and implementation of an extremely scalable real-time RFI mitigation method, based on the offline AOFlagger. All algorithms scale linearly in the number of samples. We describe how we implemented the flagger in the LOFAR real-time pipeline, on both CPUs and GPUs. Additionally, we introduce a novel simple history-based flagger that helps reduce the impact of our small window on the data. By examining an observation of a known pulsar, we demonstrate that our flagger can achieve much higher quality than a simple thresholder, even when running in real time, on a distributed system. The flagger works on visibility data, but also on raw voltages, and beam formed data. The algorithms are scale-invariant, and work on microsecond to second time scales. We are currently implementing a prototype for the time domain pipeline of the SKA central signal processor.

astro-ph.IM↗

arXiv2016

Auto-Tuning Dedispersion for Many-Core Accelerators

In this paper, we study the parallelization of the dedispersion algorithm on many-core accelerators, including GPUs from AMD and NVIDIA, and the Intel Xeon Phi. An important contribution is the computational analysis of the algorithm, from which we conclude that dedispersion is inherently memory-bound in any realistic scenario, in contrast to earlier reports. We also provide empirical proof that, even in unrealistic scenarios, hardware limitations keep the arithmetic intensity low, thus limiting performance. We exploit auto-tuning to adapt the algorithm, not only to different accelerators, but also to different observations, and even telescopes. Our experiments show how the algorithm is tuned automatically for different scenarios and how it exploits and highlights the underlying specificities of the hardware: in some observations, the tuner automatically optimizes device occupancy, while in others it optimizes memory bandwidth. We quantitatively analyze the problem space, and by comparing the results of optimal auto-tuned versions against the best performing fixed codes, we show the impact that auto-tuning has on performance, and conclude that it is statistically relevant.

cs.DC↗

arXiv2016

Real-Time Dedispersion for Fast Radio Transient Surveys, using Auto Tuning on Many-Core Accelerators

Dedispersion, the removal of deleterious smearing of impulsive signals by the interstellar matter, is one of the most intensive processing steps in any radio survey for pulsars and fast transients. We here present a study of the parallelization of this algorithm on many-core accelerators, including GPUs from AMD and NVIDIA, and the Intel Xeon Phi. We find that dedispersion is inherently memory-bound. Even in a perfect scenario, hardware limitations keep the arithmetic intensity low, thus limiting performance. We next exploit auto-tuning to adapt dedispersion to different accelerators, observations, and even telescopes. We demonstrate that the optimal settings differ between observational setups, and that auto-tuning significantly improves performance. This impacts time-domain surveys from Apertif to SKA.

astro-ph.IM↗