What we do
We are a computer science research group at the University of Torino, with scientific roots in parallel and distributed computing. Across different domains, we focus on the recurring challenges of modern computing systems: performance, scalability, programmability, reliability, security and energy efficiency.
Our current research includes:
HPC and heterogeneous computing · AI at scale and GenAI · Federated learning · Scientific workflows · Data analytics · RISC-V and emerging processors · Quantum and hybrid computing
Our approach
We combine foundational computer science with software development and experimental systems research.
Applications are not only use cases: they expose new computational challenges and help us develop better programming models, runtimes, workflows, and system software.
This is increasingly true also for modern AI. Training and inference at scale, and even more compound AI systems (e.g., agentic, RAG) raise systems challenges involving parallelism, communication, memory hierarchy, data movement, heterogeneous accelerators, resource management, reliability, security, and energy efficiency. We study these problems from a computer systems perspective, including distributed and federated learning, large-scale AI infrastructures, the execution of generative AI workloads, and the hardening of system-of-systems.
Our work is interdisciplinary and application-driven, but our core competence remains the design and understanding of parallel and distributed systems.
Work with us
We collaborate with universities, research centres, industry, and application domains where computing at scale creates new scientific or engineering challenges.
We welcome MSc student thesis, PhD candidates, postdoctoral researchers, visiting researchers, and scientific collaborators.
We proudly run HPC4AI
We designed, engineered, and operate HPC4AI, the cloud-HPC research infrastructure of the University of Torino. Built in-house rather than acquired as a turnkey system, it is engineered for high energy efficiency and reliability, with PUE < 1.1 and MTBF > 3 years. For us, computing infrastructure is also a research instrument: a place to experiment with HPC, GPUs, cloud technologies, heterogeneous systems, Arm and RISC-V platforms, and the software stack that connects them.
But research is about people before machines. Our Software & Integration Living Lab was designed as a human-centred environment for researchers, students, visitors, and collaborators working with us. It combines spaces for focused work with areas for collaboration, informal discussion, and co-creation, with particular attention to wellbeing, inclusion, and environmental quality.
Take a look at our Software & Integration Living Lab.