Data-Centric Computing Paradigm Shift, and Domain-Specific Architecture and Hardware

Memory has been an important building block of computing. Starting from Williams-Kilburn tube, the first high-speed, entirely electronic memory demonstrated in 1947, to magnetic drum memory in 1950 by US Atlas to today’s many memory technologies. Memory role is to store data which will be used by other units such as the processor or input/output. Memory architecture and design has been a critical task for achieving large storage, low latency, fast access time, and energy efficiency especially for battery-operated devices. The increased amount of data generated by many devices such as mobile, sensors, communications, and security not only increased the requirements on memory capacity, but also introduces challenges on memory access. The memory interface has limited throughput and high latency which has not been scaling at the same rate as data size or processing speed, this limits the performance of accessing the data which refer to as memory wall. In addition to negative impact on latency and performance large data movement results in high energy consumption. And due to the paramount significance of the memory block, this book focuses on the shifting in computing paradigm that is adopting various volatile and non-volatile memory technologies to perform what is known by in-memory computing.