JSSC 2024第9期Memory55nmCIM

A Compute-in-Memory Annealing Processor With Interaction Coefficient Reuse and S

提出了一种基于存内计算的数字退火处理器，用于高效解决组合优化问题。

55nm CMOS, 0.9V, 2.4fJ/spin, 402µm²/spin

存内计算退火处理器组合优化伊辛模型能效优化

▸创新点1：存内计算架构（方法创新） - 采用完全数字化的存内计算架构，将计算单元与存储单元紧密结合，显著减少数据移动带来的能量消耗和延迟，实测能效达2.4 fJ/spin@0.9V。

▸创新点2：交互系数重用策略（电路创新） - 通过设计CIM系数阵列并复用交互系数，减少硬件资源重复计算，提升面积效率至402 µm²/spin，适用于大规模组合优化问题。

▸创新点3：稀疏感知加法树（系统创新） - 动态识别稀疏计算场景，跳过无效加法操作，降低动态功耗，结合非线性概率翻转电路提升整体能效比。

▸创新点4：非线性概率翻转电路（电路创新） - 基于投票机制和片上随机数生成器，实现低硬件开销的近似计算，优化伊辛模型能量状态搜索过程，支持max-cut和图像分割等实际应用验证。

Abstract

Since combinatorial optimization problems (COPs) are a class of non-deterministic polynomial-time (NP)-hard prob- lems, it is impracticable to solve them in brute-force searches, which results in high energy consumption and long computation latency. The annealing processors based on the Ising model are naturally oriented to find approximate solutions. However, these processors face the challenges of frequent data movement between computing elements and memory units, resulting in significantly la