818- TOPS/W CSNR-31dB SQNR-45dB 10位电容器-可重构内存计算宏与Transformers的软件-模拟协同设计
An 818-TOPS/W CSNR-31dB SQNR-45dB 10-bit Capacitor-Reconfiguring Computing-in-Memory Macro with Software-Analog Co-Design for Transformers
摘要：
Transformer推理需要高计算精度;由于固有的计算误差，使用模拟CIM很难实现这一点。为了克服这一挑战，我们提出了一个电容重新配置CIM（CR-CIM），实现高计算精度的模拟CIM与10位ADC实现高面积/功率效率。CR-CIM重新配置其电容器阵列以实现双重目的：用于计算和ADC转换，从而显著节省面积。此外，CR-CIM通过在操作期间保持信号电荷稳定来消除信号衰减，从而使比较器能效提高4倍。我们还提出了一种软件模拟协同设计技术，将多数表决集成到10位ADC中，以基于运行层动态优化CIM噪声性能，从而进一步节省推理能力。我们的CR-CIM实现了模拟CIM的最高计算精度，818 TOPS/W的功率效率与最先进的竞争力。此外，考虑SQNR和CSNR的FoM分别比以前的作品高2.3倍和1.5倍。实现了视觉Transformer（ViT）推理，并为模拟CIM实现了95.8%的最高CIFAR 10精度。
英文摘要：
Transformer inference requires high compute accuracy; achieving this using analog CIMs has been difficult due to inherent computational errors. To overcome this challenge, we propose a Capacitor-Reconfiguring CIM (CR-CIM) to realize high compute accuracy analog CIM with a 10-bit ADC attaining high-area/power efficiency. CR-CIM reconfigures its capacitor array to serve dual purposes: for computation and ADC conversion, achieving significant area savings. Furthermore, CR-CIMs eliminate signal attenuation by keeping the signal charge stationary during operation, leading to a 4x improvement in comparator energy efficiency. We also propose a software-analog co-design technique integrating majority voting into the 10-bit ADC to dynamically optimize the CIM noise performance based on the running layer to further save inference power. Our CR-CIM achieves the highest compute-accuracy for analog CIMs, and the power efficiency of 818 TOPS/W is competitive with the state-of-the-art. Furthermore, the FoM considering SQNR and CSNR is 2.3x and 1.5x better than previous works, respectively. Vision Transformer (ViT) inference is achieved and realizes a highest CIFAR10 accuracy of 95.8% for analog CIMs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
