Professor Mikel Lujan, University of Manchester

The forecast for Internet-of-Things (IoT) /Edge growth expects a world with a trillion connected devices by 2030, all gathering data and triggering distributed processing pipelines that aim to minimize data movement (the most energy consuming part). Ultimately, IoT/Edge computing systems will need to analyse/filter more data, leading to increased demands on their computational capacity whilst maintaining their power envelopes. Moving forward, thermal constraints are forcing computer architectures (System-onChips) to include hardware accelerators (such as GPUs that are hardware accelerators for graphics). The UK has an almost unique opportunity to play a leading role in the development of low-power System-on-Chips (SoCs) for IoT/Edge systems. However, designing these SoCs with hardware accelerators is still highly human labour intensive and lacks appropriate integrated methodologies and tools.