Research

Unlocking the Future of Computing: CAES Pioneers Innovative Architectures to Tackle Tomorrow's Challenges.

CAES focuses on advancing the field of computer architecture and computing systems through innovative research. We are dedicated to designing, prototyping, and evaluating novel computer systems and computing paradigms across a range of scales, from embedded systems to data centres and supercomputers. With an emphasis on addressing the complexities introduced by distributed applications, such as the Internet-of-Things (IoT) and machine learning, the group's core topics include novel compute paradigms, domain-specific computer systems, dependable computing, real-time computing, and sustainable computing solutions. By striving to create architectures that meet performance, efficiency, reliability, and security demands, CAES aims to contribute to solving significant societal challenges, including energy transition, responsible AI, personalized health, and space exploration. The interdisciplinary approach of CAES combines expertise from computer science, electrical engineering, and mathematics, fostering collaboration and innovation to develop impactful technologies and methodologies.

Energy In Twente

Fifteen years ago, CAES established the ‘Energy Management’ (EM@UT) direction to advance smart grid research. EM@UT now encompasses technical, economic, and social aspects, collaborating with multiple faculties and engaging in over 10 national and international projects, ranging from conceptual to practical applications.

EM@UT webpage

Dependable Computing Systems

The Dependable Computing Systems Interest Group focuses on creating reliable and trustworthy computing systems that can withstand failures. Their research encompasses fault tolerance, system reliability, and security, addressing the complexities of modern computing. The team collaborates with academic and industry partners, aiming to deliver innovative solutions that enhance the robustness and security of computing systems for real-world applications.

Discover more about DCS!

Brain-Inspired Processing Systems

Neuromorphic research at CAES centers on developing scalable, brain-inspired processing systems that provide energy-efficient computation to meet the needs of large-scale AI algorithms. This involves investigating event-driven, memory-centric compute fabrics that reduce data movement for sparse, adaptive execution, as well as designing neuromorphic algorithms alongside hardware. This co-design optimizes model structures, scheduling, mapping, and on-device learning rules to align with processor primitives.

Read more

Visit the individual focus groups for more information.