Foto del docente

Andrea Bartolini

Senior assistant professor (fixed-term)

Department of Electrical, Electronic, and Information Engineering "Guglielmo Marconi"

Academic discipline: ING-INF/05 Information Processing Systems

Curriculum vitae

Download Curriculum Vitae (.pdf 137KB )

Andrea Bartolini is Junior assistant professor (fixed-term) of Electronic Engineering at University of Bologna, Department of Electrical, Electronic, and Information Engineering "Guglielmo Marconi". In 2017, he received the Scientific National Qualification in Italy as Associate Professor in sector 09/H1 Information Processing Systems. Since 2007, Andrea Bartolini has published (international, peer-reviewed) 22 journal papers and 64 conference papers. Publications includes top IEEE and ACM journals.

He started his scientific research activity at Freescale Semiconductor LTD with research activity on multimedia embedded platforms, optimization of the visual perception of reproduced videos under low power strategies. During this period, he became a Linux kernel hacker and low-level programmer of Linux based embedded systems.

During his PhD he expanded this research topic on cross-domain (embedded and high-performance computing) multicore and many-core platforms, increasing his knowledge in architecture level performance, power, and thermal modelling. Due to the complexity of modelling the link of these physical effects with the workload running on the multi-core he trained on machine learning and system identification achieving important scientific contributions such as regression approaches to thermal sensor calibration, robust bias compensated autoregressive models, artificial neural networks, and regression trees for power and thermal model extraction.

During his stay at Intel Lab Germany he had the chance to gain a deeper understanding of Intel server platforms and the many-core system SW stack, including the Linux O.S. power management and monitoring subsystems. He was an active member of the Intel Single Chip Cloud Computing community named MARC.

During his post-doctoral research activities at University of Bologna (UNIBO), he was involved in the preparation and implementation of the PHIDIAS FP7 FET project in which he had the opportunity of learning advance knowledge in ultra-low power design, body area sensor networks, and advance signal processing for biomedical signals. He was involved in the preparation and currently active implementation of the MULTITHERMAN ERC Advance project in which he led the activity on thermal and energy efficient design for high-performance computing systems. During the same period, he started a fruitful collaboration with CINECA - the Italian supercomputer centre (hosting a Tier0 PRACE supercomputer as well as the greenest supercomputer in the world Eurora in November 2013). Thanks to this experience he has gained in-depth knowledge in supercomputer design and challenges. Between 2013 and early 2017, he worked as post-doctoral researcher in the Integrated System Institute in the D-ITET department at ETH Zurich, while maintaining a part-time collaboration with the DEI department in UNIBO. In Zurich, he started working in two Nanotera projects, the YINS project for energy-efficient server platforms based on ultra-low power electronic techniques, as well as in the Ultrasound2Go where he studied hardware accelerated beam forming algorithms for 3D imaging. Since 2015 he has been involved as work package leader in the ANTAREX project and as technical contributor in the EXANODE project with focus on HPC system monitoring and next Exascale node architecture. He is an expert member in the ETP4HPC for the definition of the European HPC agenda, and he has been involved as a consultant for E4 Engineering with responsibility in the energy-efficient design of future green HPC nodes. During the work with E4 Engineering he has specialized in the design of embedded systems for the fine grain power and temperature monitoring based on open source hardware. As the result of the collaboration with E4 Engineering the developed D.A.V.I.D.E. supercomputer ranked in July and November in the top twenty most energy efficient supercomputers worldwide.

Based on Google Scholar (23/03/2018), his h-index is 15, his i10-index is 26 and his total number of citations is 829. Based on Scopus (23/03/2018) his h-index is 13, and his total number citations is 602.