Computing-in-memory (CiM) is a promising technique to achieve high energy efficiency in data-intensive matrix-vector multiplication (MVM) by relieving the memory bottleneck. Unfortunately, due to the limited SRAM capacity, existing SRAM-based CiM needs to reload the weights from DRAM in large-scale networks. This undesired fact weakens the energy efficiency significantly. This work, for the first time, proposes the concept, design, and optimization of computing-in-ROM to achieve much higher on-chip memory capacity, and thus less DRAM access and lower energy consumption. Furthermore, to support different computing scenarios with varying weights, a weight fine-tune technique, namely Residual Branch (ReBranch), is also proposed. ReBranch combines ROM-CiM and assisting SRAM-CiM to ahieve high versatility. YOLoC, a ReBranch-assisted ROM-CiM framework for object detection is presented and evaluated. With the same area in 28nm CMOS, YOLoC for several datasets has shown significant energy efficiency improvement by 14.8x for YOLO (Darknet-19) and 4.8x for ResNet-18, with <8% latency overhead and almost no mean average precision (mAP) loss (-0.5% ~ +0.2%), compared with the fully SRAM-based CiM.
翻译:电子计算机化( CiM) 是一个很有希望的技术, 通过缓解内存瓶颈, 在数据密集型矩阵矢量倍增( MVM) 中实现高能效, 是一个很有希望的技术。 不幸的是, 由于 SRAM 能力有限, 以 SRAM 为基础的 SRAM 现有 CIM 需要在大型网络中重新装载 DRAM 的重量。 这个不理想的事实大大削弱了能源效率。 这项工作首次提出了计算在轨存储能力的概念、 设计和优化, 从而降低 DRAM 访问量和能源消耗量。 此外, 支持不同计算情景, 以不同重量支持不同的计算情景, 还提出了一种重量微调技术, 即 残留处( Rebrach ) 。 ReBranch 将 ROM- CiM 和 协助 SRAM- CiM 到高多功能。 YOLoC 首次提出并评估了基于 ReBranch 的 ROM- CiM 框架( 完全用于目标检测, ROM- CiM 框架 ), 和 14 MOS, YOL- DOC 和 中位 ( YOL- DRVER ) 等 平均能量 改进了能源效率, 14 和 的YOL- 14% 和 和 和 等 中 等 等 中度 中度 中度 等 中 等 等 等 的能量 损耗损耗损率 。