Biography
Pedro Tomás is an Associate Professor at the Department of Electrical and Computer Engineering of IST and a Senior Researcher at Instituto de Engenharia de Sistemas e Computadores R&D (INESC-ID). His research is focused on general-purpose and specialized processor architectures (e.g., CPUs and GPUs) and reconfigurable and heterogeneous systems (e.g., based on FPGAs) for embedded and high-performance computing systems. He has been developing techniques for energy-efficient optimization of computing platforms, which can be achieved at the architecture level, by specifically adapting the architecture to the target application (e.g., by relying on reconfigurable computing approaches), and at the software or Operating System (OS) levels, by developing energy-aware strategies for thread scheduling and dynamic voltage and frequency scaling of CPUs and GPUs. He is also working on new stream-based instruction-set architectures and extensions, such as by relying on efficient memory access pattern descriptions, to promote timely and effective prefetching of data in intensive kernels. He has recently been interested in deep learning approaches for multiple domains (e.g., for the performance, power and energy modelling of processor architectures; for video coding; for disease diagnosis; or for clustering).
Pedro Tomás is a Senior Member of IEEE, currently affiliated with the Computer and Signal Processing Societies, and has participated in the reviewing process of multiple EU projects.
Main Interests
High-Performance Computing, Computer Architectures, Parallel Computing, Energy-Aware and Power-Constrained Architectures and SystemsHigh-Performance Computing, Computer Architectures, Parallel Computing, Energy-Aware and Power-Constrained Architectures and Systems