BWBen van Werkhoven

Ben van Werkhoven

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Ben van Werkhoven has a total of 132 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Velvet: Parallel Divide-and-Conquer in Safe Rust

    by Badia Liokouras, Ben van Werkhoven, Rob V. van Nieuwpoort

  • Open Access

    Research Software Engineers in the Age of GenAI: Same Value, Changing Practice

    by Stephan Druskat, Michelle Barker, Ian Cosden, Cunliang Geng, Robert Haines, Daniel S. Katz, Joseph Shingleton, Ben van Werkhoven

    Authors: Stephan Druskat, Michelle Barker, Ian Cosden, Cunliang Geng, Robert Haines, Daniel S. Katz, Joseph Shingleton, Ben van Werkhoven 10.5281/zenodo.20320178 Generative AI, specifically Large Language Models (LLMs) and coding agents, is rapidly changing software engineering.

  • Open Access

    Research Software Engineers in the Age of GenAI: Same Value, Changing Practice

    by Stephan Druskat, Michelle Barker, Ian Cosden, Cunliang Geng, Robert Haines, Daniel S. Katz, Joseph Shingleton, Ben van Werkhoven

    Authors: Stephan Druskat, Michelle Barker, Ian Cosden, Cunliang Geng, Robert Haines, Daniel S. Katz, Joseph Shingleton, Ben van Werkhoven 10.5281/zenodo.20320178 Generative AI, specifically Large Language Models (LLMs) and coding agents, is rapidly changing software engineering.

  • Open Access

    Kernel Float: Unlocking Mixed-Precision GPU Programming

    by Stijn Heldens, Ben van Werkhoven

    Modern GPUs feature specialized hardware for low-precision floating-point arithmetic to accelerate compute-intensive workloads that do not require high numerical accuracy, such as those from artificial intelligence. However, despite the significant gains in computational throughput, memory bandwidth utilization, and energy efficiency, integrating low-precision formats into scientific applications remains difficult. We introduce Kernel Float , a header-only C++ library that simplifies the development of portable mixed-precision GPU kernels. Kernel Float provides a generic vector type, a unified interface for common mathematical operations, and fast approximations for low-precision transcendental functions that lack native hardware support. To demonstrate the potential of mixed-precision computing unlocked by our library, we integrated Kernel Float into nine GPU kernels from various domains. Our evaluation on Nvidia A100 and AMD MI250X GPUs shows performance improvements of up to \(12\times\) over double precision, while reducing source code length by up to 50% compared to handwritten kernels and having negligible runtime overhead. Our results further show that mixed-precision performance depends not only on choosing appropriate data types but also on tuning traditional optimization parameters (e.g., block size and vector width) and, when relevant, even domain-specific parameters.


  • Open Access

    Accuracy-Aware Mixed-Precision GPU Auto-Tuning

    by Stijn Heldens, Ben van Werkhoven

    Reduced-precision floating-point arithmetic has become increasingly important in GPU applications for AI and HPC, as it can deliver substantial speedups while reducing energy consumption and memory footprint. However, choosing the appropriate data formats brings a challenging tuning problem: precision parameters must be chosen to maximize performance while preserving numerical accuracy. At the same time, GPU kernels typically expose additional tunable optimization parameters, such as block size, tiling strategy, and vector width. The combination of these two kinds of parameters results in a complex trade-off between accuracy and performance, making manual exploration of the resulting design space time-consuming. In this work, we present an accuracy-aware extension to the open-source Kernel Tuner framework, enabling automatic tuning of floating-point precision parameters alongside conventional code-optimization parameters. We evaluate our accuracy-aware tuning solution on both Nvidia and AMD GPUs using a variety of kernels. Our results show speedups of up to $12{\times }$ over double precision, demonstrate how Kernel Tuner’s built-in search strategies are effective for accuracy-aware tuning, and show that our approach can be extended to other optimization objectives, such as memory footprint or energy efficiency. Moreover, we highlight that jointly tuning accuracy- and performance-affecting parameters outperforms isolated approaches in finding the best-performing configurations, despite significantly expanding the optimization space. This unified approach enables developers to trade accuracy for throughput systematically, enabling broader adoption of mixed-precision computing in scientific and industrial applications.


  • Open Access

    Efficient Construction of Large Search Spaces for Auto-Tuning

    by Floris-Jan Willemsen, Rob V. van Nieuwpoort, Ben van Werkhoven


  • Open Access

    PowerSensor3: A Fast and Accurate Open Source Power Measurement Tool

    by Steven van der Vlugt, Leon Oostrum, Gijs Schoonderbeek, Ben van Werkhoven, Bram Veenboer, Krijn Doekemeijer, John W. Romein

    Power consumption is a major concern in data centers and HPC applications, with GPUs typically accounting for more than half of system power usage. While accurate power measurement tools are crucial for optimizing the energy efficiency of (GPU) applications, both built-in power sensors as well as state-of-the-art power meters often lack the accuracy and temporal granularity needed, or are impractical to use. Released as open hardware, firmware, and software, PowerSensor3 provides a cost-effective solution for evaluating energy efficiency, enabling advancements in sustainable computing. The toolkit consists of a baseboard with a variety of sensor modules accompanied by host libraries with C++ and Python bindings. PowerSensor3 enables real-time power measurements of SoC boards and PCIe cards, including GPUs, FPGAs, NICs, SSDs, and domain-specific AI and ML accelerators. Additionally, it provides significant improvements over previous tools, such as a robust and modular design, current sensors resistant to external interference, simplified calibration, and a sampling rate up to 20 kHz, which is essential to identify GPU behavior at high temporal granularity. This work describes the toolkit design, evaluates its performance characteristics, and shows several use cases (GPUs, NVIDIA Jetson AGX Orin, and SSD), demonstrating PowerSensor3's potential to significantly enhance energy efficiency in modern computing environments.

  • Open Access

    Calibration-free estimation of field dependent aberrations for single molecule localization microscopy across large fields of view

    by Isabel Droste, Erik Schuitema, Sajjad Khan, Stijn Heldens, Ben van Werkhoven, Keith A. Lidke, Sjoerd Stallinga, Bernd Rieger

    Image quality in single molecule localization microscopy (SMLM) depends largely on the accuracy and precision of the localizations. While under ideal imaging conditions the theoretically obtainable precision and accuracy are achieved, in practice this changes if (field dependent) aberrations are present. Currently there is no simple way to measure and incorporate these aberrations into the Point Spread Function (PSF) fitting, therefore the aberrations are often taken constant or neglected all together. Here we introduce a model-based approach to estimate the field-dependent aberration directly from single molecule data without a calibration step. This is made possible by using nodal aberration theory to incorporate the field-dependency of aberrations into our fully vectorial PSF model. This results in a limited set of aberration fit parameters that can be extracted from the raw frames without a bead calibration measurement, also in retrospect. The software implementation is computationally efficient, enabling fitting of a full 2D or 3D dataset within a few minutes. We demonstrate our method on 2D and 3D localization data of microtubuli and nuclear pore complexes over fields of view (FOV) of up to 180 μm and compare it with spline-based fitting and a deep learning based approach.

  • Open Access

    A methodology for comparing optimization algorithms for auto-tuning

    by Floris-Jan Willemsen, Richard Schoonhoven, Jiří Filipovič, Jacob O. Tørring, Rob van Nieuwpoort, Ben van Werkhoven