The 'Riding on Moore's Law' project
aims to initiate a new approach to supercomputer design, based on the co-design of the architecture and the software runtime environment, to overcome the current stagnation in the efficiency of computing systems. Since the mid-1960s, Moore's Law has held true, stating that the number of components on a chip doubles approximately every two years. An expected consequence is that computer performance would also double approximately every two years. This empirical law has held true, and continues to hold true today, in terms of integration capacity, but the expected performance improvement stalled in the early years of this century. Due to energy consumption issues and the high sophistication/complexity of core design, doubling the number of components used in its design resulted in a failure to double performance. The alternative has been to increase the number of cores, shifting the burden of efficient use to the programmer.

The project led by Mateo Valero proposes a radically new conception of parallel architecture computers, raising the level of abstraction offered to the programmer, who will now think in terms of tasks rather than low-level instructions. The system software will be responsible for mapping computational demands (tasks) to the system's physical resources (cores, memory).
Fundamental elements of this architecture will be vector processors and techniques, given their high energy efficiency. The project argues for (demonstrates) the need to design the architecture and system software together, so that the result of this collaboration is the optimization of the efficiency and productivity of computer systems.

Nine of the researchers who have received this prestigious grant work at centers located in Catalonia, representing 60% of the total in Spain. Thus, in addition to Valero, the new Advanced Grant recipients are:
Xavier Oliver (International Centre for Numerical Methods in Engineering), Núria Sebastián-Galles and Vicent Caselles (Pompeu Fabra University), Xavier Tolsa (Autonomous University of Barcelona), Antonio Echavarren (Catalan Institute of Chemical Research), Albert Marcet (Markets, Organizations and Votes in Economics), and Susana Narotzjy (University of Barcelona).

Brief CV of Mateo Valero
: PhD in Telecommunications Engineering. Since 1983, he has been a professor at the Polytechnic University of Catalonia (UPC). He has published more than 600 articles in the area of ​​high-performance computer architecture. Director of the Barcelona Supercomputing Center – National Supercomputing Center. Among his awards are the 2007 Eckert-Mauchly Prize, the world's most prestigious award in the field of Computer Architecture; the 2009 Harry H. Goode Award from the IEEE; two national research awards: the Julio Rey Pastor Prize in Computer Science and Mathematics and the Leonardo Torres Quevedo Prize in Engineering; the King Jaime I Research Prize from the Valencian Government; and induction into the ICT European Program's Hall of Fame, selected as one of the 25 most influential European researchers in IT (Information Technology). He is an IEEE Fellow, a Distinguished Fellow of Intel, and a Fellow of the ACM. He is a founding member of the Royal Academy of Engineering of Spain, a corresponding member of the Royal Academy of Exact, Physical and Natural Sciences, a member of the Royal Academy of Sciences and Arts of Barcelona, ​​a member of the European Academy, and a corresponding member of the Mexican Academy of Sciences. He holds honorary doctorates from the Universities of Chalmers, Belgrade, Las Palmas, Veracruz, and Zaragoza. In 1998, he was chosen as a favorite son of his hometown, Alfamén, which decided to name its school after him in 2006.

More information