This project proposes the development, integration, and physical implementation of specialized hardware accelerators to overcome the primary performance bottlenecks in Monte Carlo particle transport simulators. Focusing on the dominant cross-section (XS) lookup kernel, we will refine our novel processing-near-memory accelerator design and explore its adaptation for next-generation cross-section representation methods. Furthermore, we will investigate the integration of these accelerators within a customized RISC-V manycore system architecture, employing a hardware-software codesign approach to optimize simulator algorithms. The final deliverable will be a manufacturable GDSII layout, targeting high performance modern process node, with an option for a small-scale ASIC prototype.

Cíle udržitelného rozvoje

Masarykova univerzita se hlásí k cílům udržitelného rozvoje OSN, jejichž záměrem je do roku 2030 zlepšit podmínky a kvalitu života na naší planetě.