Project: Multicore Programming
Multicore/Multiprocessor hardware platforms are finding widespread use in virtually all embedded ICT domains, including wireless communication, smartphones, automotive, computer vision, and many others. From a pure HW perspective, multicores are the key enablers for continued performance scaling in forthcoming compute-intensive applications, while keeping power/energy consumption and power density within reasonable limits. Systems-on-Chip (SoC´s) with dozens of programmable cores are becoming commodity today, and the ITRS roadmap predicts devices with 1000+ processor cores in the coming years. However, software development for multicores becomes a significant bottleneck. Migrating legacy application code or firmware as well as developing and debugging new software for highly parallel HW platforms causes a significant embedded SW productivity gap. This lab course provides hands-on experience in embedded multicore programming using advanced parallelizing compiler technology from Silexica, an ICE spin-off company.