A compute-in-memory chip based on resistive random-access memory.
A compute-in-memory chip based on resistive random-access memory.
复制标题
DOI：
10.1038/s41586-022-04992-8
复制
发表时间：
2022-08
期刊：
影响因子：
64.8
通讯作者：
Cauwenberghs, Gert
中科院分区：
文献类型：
--
作者：
Wan, Weier;Kubendran, Rajkumar;Schaefer, Clemens;Eryilmaz, Sukru Burc;Zhang, Wenqiang;Wu, Dabin;Deiss, Stephen;Raina, Priyanka;Qian, He;Gao, Bin;Joshi, Siddharth;Wu, Huaqiang;Wong, H-S Philip;Cauwenberghs, Gert
直接在边缘设备上实现越来越复杂的人工智能（AI）功能，要求基于电阻的随机记忆（RRAM）来实现Edge硬件的前所未有的能源效率。在密集的，模拟和非挥发性RRAM设备中，并通过直接在RRAM中执行AI计算，从而消除了单独的计算和内存之间的渴望渴望的数据运动。尽管最近的研究表明，在完全集成的RRAM-CIM硬件上表明，RRAM-CIM芯片同时提供了高能量效率，以支持潜水员模型和软件可观的准确性，多功能性和准确性对于广泛采用该技术是必不可少的，与之相关的权衡不能通过通过孤立的改进来解决在这里设计的任何单一抽象水平，通过在算法和架构到电路和设备的设计的所有层次结构中进行优化，我们提出了基于RRAM的CIM芯片体系结构，能源效率比以前在各种计算位构成的先前最先进的RRAM-CIM芯片好两次，并且推理精度与各种AI任务量化的软件模型相当，包括MNIST的精度为99.0％和CIFAR-10图像分类的85.7％，Google语音命令识别的84.7％的准确性和70％的降低。在贝叶斯图像重返任务上的图像重建错误中。
与现有硬件相比，通过在设计的所有层次结构中进行优化，基于电阻随机访问存储器的计算中的神经网络推理加速器只能提高能源效率，灵活性和准确性。
Realizing increasingly complex artificial intelligence (AI) functionalities directly on edge devices calls for unprecedented energy efficiency of edge hardware. Compute-in-memory (CIM) based on resistive random-access memory (RRAM) promises to meet such demand by storing AI model weights in dense, analogue and non-volatile RRAM devices, and by performing AI computation directly within RRAM, thus eliminating power-hungry data movement between separate compute and memory. Although recent studies have demonstrated in-memory matrix-vector multiplication on fully integrated RRAM-CIM hardware, it remains a goal for a RRAM-CIM chip to simultaneously deliver high energy efficiency, versatility to support diverse models and software-comparable accuracy. Although efficiency, versatility and accuracy are all indispensable for broad adoption of the technology, the inter-related trade-offs among them cannot be addressed by isolated improvements on any single abstraction level of the design. Here, by co-optimizing across all hierarchies of the design from algorithms and architecture to circuits and devices, we present NeuRRAM—a RRAM-based CIM chip that simultaneously delivers versatility in reconfiguring CIM cores for diverse model architectures, energy efficiency that is two-times better than previous state-of-the-art RRAM-CIM chips across various computational bit-precisions, and inference accuracy comparable to software models quantized to four-bit weights across various AI tasks, including accuracy of 99.0 percent on MNIST and 85.7 percent on CIFAR-10 image classification, 84.7-percent accuracy on Google speech command recognition, and a 70-percent reduction in image-reconstruction error on a Bayesian image-recovery task. A compute-in-memory neural-network inference accelerator based on resistive random-access memory simultaneously improves energy efficiency, flexibility and accuracy compared with existing hardware by co-optimizing across all hierarchies of the design.
登录
查看更多内容
影响因子：
3.1
作者：
Burr, Geoffrey W.;Shelby, Robert M.;Hwang, Hyunsang
通讯作者：
Hwang, Hyunsang
影响因子：
5.4
作者：
Chen, Yu-Hsin;Krishna, Tushar;Sze, Vivienne
通讯作者：
Sze, Vivienne
影响因子：
7.8
作者：
HORNIK, K;STINCHCOMBE, M;WHITE, H
通讯作者：
WHITE, H
影响因子：
3.1
作者：
Eryilmaz, Sukru Burc;Neftci, Emre;Wong, Hon-Sum Philip
通讯作者：
Wong, Hon-Sum Philip
影响因子：
64.8
作者：
Jung, Seungchul;Lee, Hyungwoo;Kim, Sang Joon
通讯作者：
Kim, Sang Joon