We are pleased to announce the winner of the ISC 2026 Hans Meuer Award for the best research paper of the year. The winning paper was selected from 122 full submissions, underscoring the competitiveness of our paper program.The Hans Meuer Award, which comes with a cash prize of €3,000, has been awarded to  “PICO: Performance Insights for Collective Operations,” authored by Saverio Pasqualoni from King Abdullah University of Science and Technology (KAUST) and Sapienza University of Rome, Tommaso Bonato and Torsten Hoefler from ETH Zürich, Lorenzo Piarulli and Daniele De Sensi from Sapienza University of Rome, and Marco Canini from Kaust.

The paper presents PICO, an open-source framework designed for analyzing and optimizing collective communication operations, which are a significant performance bottleneck in large-scale HPC and AI systems. By facilitating reproducible benchmarking and detailed performance analysis across hardware and software layers, PICO reveals that the default communication settings can be up to five times slower than the optimal configurations. Furthermore, the framework demonstrates that targeted tuning can reduce AI training times by up to 44%, offering practical insights to improve efficiency in next-gen computing environments.

This year, a total of 13 papers have been selected for presentation at the conference. All accepted papers will be published as open access in the IEEE Xplore Digital Library. The research paper committee for 2026 is chaired by Hatem Ltaief from Kaust, with Richard Vuduc from the Georgia Institute of Technology serving as co-chair.

Best Student Paper Award
The inaugural Best Student Paper Award has been presented to “Energy Efficiency in Analog Photonic Processors: Conversions and Losses at Scale.” This paper was authored by Niklas Bahr from Volkswagen group innovation and Heidelberg university, Daniel Steinmeyer from Volkswagen group innovation, and Wolfram Pernice from Heidelberg university.

This paper introduces a unified analytical framework for assessing the energy efficiency of analog photonic processors, which are a promising alternative to electronic computing for AI workloads. The study emphasizes the critical role of signal conversion and system-level losses, showing that while optical computation is inherently efficient, overall performance depends heavily on architecture and scaling effects. The authors identify specific conditions under which photonic systems can outperform conventional hardware and propose a hybrid electro-optical design to improve scalability and efficiency.

The Best Student Paper Award includes a €1,000 cash prize, generously sponsored by Jeff Hammond, distinguished engineer, Nvidia.

Both winning papers will be presented in person at ISC High Performance 2026:

• Hans Meuer Award – Wednesday, June 24, 11:15-11:45 am
• Best Student Research Paper Award – Tuesday, June 23, 5:15-5:45 pm

Read also :