Conventional computing architectures have recently witnessed unprecendented challenges as emerging applications demand more computation, and process increasingly larger datasets. Advancements in machine learning, computer vision, data mining, graph processing, and data analytics have stressed conventional multicore, CPU-centric architectures to the point where their efficacy at handling these workloads has been called in to question. At the same time, new technological opportunities and trends such as die stacked memory, server side FPGA offload, near data processing, and higher density non-volatile mass storage have opened up an enormous architectural research space to support these new developing applications. This project aims to explore the space of near data processing architectures in the context of next generation mass storage devices for next generation server machines.
