2021/08/17 by Weier Wan, Rajkumar Kubendran, Wan, Weier +27
Computer Science · Engineering · #Advanced Memory and Neural Computing #Advanced Neural Network Applications #Artificial Intelligence (cs.AI) #Emerging Technologies (cs.ET) #FOS: Computer and information sciences #Ferroelectric and Negative Capacitance Devices #Hardware Architecture (cs.AR) #Machine Learning (cs.LG) #cs.AI #cs.AR #cs.ET #cs.LG
paper · pdf · doi:10.48550/arxiv.2108.07879
34 pages, 14 figures, 1 table
arxiv created 2021/08/17 · openalex publication_date 2021/08/17 · arxiv updated 2021/08/19 · openalex created_date 2021/08/30 · openalex updated_date 2026/07/28
Realizing today's cloud-level artificial intelligence functionalities directly on devices distributed at the edge of the internet calls for edge hardware capable of processing multiple modalities of sensory data (e.g. video, audio) at unprecedented energy-efficiency. AI hardware architectures today cannot meet the demand due to a fundamental "memory wall": data movement between separate compute and memory units consumes large energy and incurs long latency. Resistive random-access memory (RRAM) based compute-in-memory (CIM) architectures promise to bring orders of magnitude energy-efficiency improvement by performing computation directly within memory. However, conventional approaches to CIM hardware design limit its functional flexibility necessary for processing diverse AI workloads, and must overcome hardware imperfections that degrade inference accuracy. Such trade-offs between efficiency, versatility and accuracy cannot be addressed by isolated improvements on any single level of the design. By co-optimizing across all hierarchies of the design from algorithms and architecture to circuits and devices, we present NeuRRAM - the first multimodal edge AI chip using RRAM CIM to simultaneously deliver a high degree of versatility for diverse model architectures, record energy-efficiency 5× - 8× better than prior art across various computational bit-precisions, and inference accuracy comparable to software models with 4-bit weights on all measured standard AI benchmarks including accuracy of 99.0% on MNIST and 85.7% on CIFAR-10 image classification, 84.7% accuracy on Google speech command recognition, and a 70% reduction in image reconstruction error on a Bayesian image recovery task. This work paves a way towards building highly efficient and reconfigurable edge AI hardware platforms for the more demanding and heterogeneous AI applications of the future.