发表机构
University of Minnesota; Case Western Reserve University; University of Louisiana at Lafayette（明尼苏达大学; 凯斯西储大学; 路易斯安那拉斐特大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对边缘AI应用的能效与可靠性需求，提出基于MTJ和随机计算的容错存内架构FALCON，经14 nm FinFET工艺实验验证其在严苛条件下仍具备正确功能，可用于可靠性关键型边缘AI场景。
AI 中文摘要
随着神经推理和传感器边缘分析等现代数据中心应用的扩展，它们日益遭遇冯·诺依曼内存墙问题，承受着过高的数据移动开销和严格的能耗约束。利用磁隧道结（MTJ）等新型非易失性技术的存内计算（IMC）有望缓解这些瓶颈。然而，传统基于二进制基数的IMC架构易受工艺诱导的变异影响，存在受限的工作裕度和热噪声问题。为弥合能效与计算可靠性之间的差距，本研究提出FALCON，一种结合随机计算（SC）的、基于MTJ的容错存内算术架构。通过将数值编码为均匀比特流，SC可自然吸收局部软错误，并能利用高度紧凑的逻辑原语在存储阵列内执行基本算术运算套件。FALCON将确定性比特映射机制与可重构存内逻辑（LIM）结构相集成，无需将数据传输至外部处理器，也无需依赖面积和功耗高昂的随机数生成器。采用14 nm FinFET工艺的实验结果验证，即便在激进的电压缩放、严重的工艺变异以及高达30%的噪声注入水平下，FALCON仍具备正确功能，使其成为面向可靠性关键型边缘AI应用的稳健框架。本研究还以形态学闭运算作为实际的容错图像处理案例，验证了FALCON的正确功能。
英文摘要
As modern data-centric applications such as neural inference and sensor-edge analytics expand, they increasingly encounter the von Neumann memory wall, suffering from excessive data movement overhead and stringent energy constraints. In-Memory Computing (IMC) utilizing emerging non-volatile technologies, such as Magnetic Tunnel Junctions (MTJs), promises to mitigate these bottlenecks. However, conventional binary radix-based IMC architectures suffer from excessive vulnerability to process-induced variations, restricted operating margins, and thermal noise. To bridge the gap between energy efficiency and computational reliability, this work proposes FALCON, a fault-tolerant, MTJ-based in-memory arithmetic architecture integrated with Stochastic Computing (SC). By encoding numerical values into uniform bit-streams, SC naturally absorbs localized soft errors and enables the execution of an essential suite of arithmetic operations using highly compact logic primitives directly within the memory arrays. FALCON integrates a deterministic bit mapping mechanism with reconfigurable logic-in-memory (LIM) structures, eliminating the need to transfer data to external processors or area- and power-hungry random number generators. Experimental results using 14 nm FinFET technology validate the correct functionality of FALCON even under aggressive voltage scaling, severe process variation, and noise injection levels up to 30%, making it a robust framework for reliability-critical edge AI applications. We investigate the proper functionality of FALCON on morphological closing as a realistic noise-tolerant image processing case study.
