Project Description
The student within this project will be trained in programming for heterogeneous computing architectures, and in simultaneously optimizing data formats and processing techniques to enable hybrid architectures to work together to solve problems which none of these technologies could solve on their own, in partnership with a number of computing hardware companies. The main focus of the PhD will be on CPU and GPU systems, as that is the chosen solution for the LHCb collaboration’s trigger. Experts at LPNHE will contribute to enhance the PhD training in the FPGA aspect. The PhD student will work on a novel ML method for optimizing heterogeneous computing architectures, and deploy it in the context of the specific requirements of real-time data processing in the LHC collaborations, in collaboration with the other students based in the Paris area. This optimization code will then be applied to the specific problems of the LHCb and ATLAS real-time data-processing architectures. The expertise in hybrid architectures and ML be employed in commercial collaborations between CERN and self-driving car companies, and on traffic predictions in collaboration with Ximantis.
