Energy-efficient In-Memory Address Calculation

Computation-in-Memory (CIM) is an emerging computing paradigm to address memory bottleneck challenges in computer architecture. A CIM unit cannot fully replace a general-purpose processor. Still, it significantly reduces the amount of data transfer between traditional and processor by enriching transferred information. Data transactions consist access addresses values. While main focus field in-memory apply computations on content (values), importance CPU-CIM calculations for generating sequence data-dominated applications generally overlooked. However, information used “address” can easily be even more than half total bits many applications. In this article, we propose circuit perform Address Calculation Accelerator. Our simulation results showed that calculating sequences inside (instead CPU) reduce therefore contribute considerable energy saving, latency, bus traffic. For chosen application guided image filtering, calculation almost two orders magnitude reduction over bus.