Posts

I am happy to share that our paper, “Instruction Error Detection with Bloom Filters: Architecture, Radiation Testing and Fault Injection,” has been accepted by IEEE Transactions on Device and Materials Reliability.

Nice news from our group: the work of my PhD student Jeffrey Spaan, presented at Computing Frontiers 2026, was recently covered by IEEE Spectrum. The paper studies how dynamic voltage and frequency scaling can be used during LLM training to reduce GPU energy consumption.

I am pleased to share that I will serve as Topic Chair for A6 – Applications of Artificial Intelligence Systems at DATE 2027. DATE has been one of the main venues where I have presented and followed research on embedded systems, computer architecture, dependable computing, and design automation over the years.

Last week, we ran the ASCI A3 course “Design and Implementation of Real-Time Systems”. It was a very enjoyable week with PhD students from different areas of computer science, including real-time systems, GPU computing, cloud and distributed systems, and resource management.

I am happy to share that our NWO–JST project, “Cross-Paradigm Orchestration of Quantum Computers, Neuromorphic Chips, and Conventional Accelerators,” has been awarded funding under the joint NWO–JST call for unconventional information processing technologies.

In 2024, we launched the first edition of the ASCI doctoral course Design and Implementation of Real-Time Systems. It brought together PhD students from different Dutch universities and created a very rewarding week of lectures, hands-on exercises, and discussions on real-time systems research and practice.

I am pleased to share that we are organizing the ACM Europe & Dutch 4TU.NIRICT Summer School on “Systems Meet AI”, which will take place in Delft, the Netherlands, from July 6–10, 2026.

Congratulations to Ivy Zhuang (Santa Clara University) for being accepted to Google Summer of Code 2026 with the RTEMS project! I will be mentoring Ivy together with Gedare Bloom on a project that brings MPCP and DPCP—two classical multiprocessor synchronization protocols—into RTEMS.

I’m happy to share that our paper “How Much Energy Is Wasted in LLM Operations? Evidence from Kernel-Level DVFS” has been accepted at the ACM International Conference on Computing Frontiers.

I’m excited to share that our paper “Efficient Hash-to-Index via Rejection Sampling for Online Fault Detection with Bloom/Cuckoo Filters” has been accepted at ETS 2026! In this work, we revisit a classic idea — rejection sampling — and show how it can be used as a practical, hardware-efficient alternative to division or multiplication for mapping hashes to table indices.