TensorCIM: A 28nm 3.7nJ/Gather and 8.3TFLOPS/W FP32 Digital-CIM Tensor Processor for MCM-CIM-Based Beyond-NN Acceleration

📋 论文概要

本文提出TensorCIM，一款基于28nm工艺的数字CIM张量处理器，用于加速推荐模型等超越神经网络的稀疏聚集和稀疏代数操作，通过MCM-CIM架构有效缓解数据移动瓶颈。

💡 主要创新点

• 提出数字CIM架构，专门针对稀疏聚集(SpG)和稀疏代数(SpA)操作进行优化，实现高效稀疏张量处理。
• 采用多芯片模块(MCM)集成CIM，支持大规模稀疏计算，并实现3.7nJ/Gather的能效和8.3TFLOPS/W的FP32算力。

🏷 关键词

📄 原文摘要

Recommendation Models (DLRMs) have computational and data-movement requirements beyond those seen in typical NN processing. Such beyond-NN applications typically consist of Sparse Gathering (SpG) and Sparse Algebra (SpA). SpG comprises gathering and reducing tensors from sparsely distributed addresses (in GCN’s aggregation phase and DLRM’s embedding layer). SpA refers to NN-based sparse tensor multiplication for the gathered tensors (in GCN’s combination phase and DLRM’s fullyconnected layer). Due to the large application size, data movement is the main bottleneck for beyond-NN acceleration. Digital Computing-In-Memory (CIM) is an efficient and precise architecture for reducing data movement [1-3]. Large-scale beyond-NN acceleration motivates the demand for scaling out digital CIM processors. However, a

👥 作者与机构

Fengbin Tu, Yiqi Wang, Zihan Wu, Weiwei Wu, Leibo Liu, Yang Hu,

Shaojun Wei, Shouyi Yin Tsinghua University, Beijing, China Applications such as Graph Convolutional Networks (GCNs) and Deep Learning