ChewBaccaNN：一种灵活的223 TOPS/W BNN加速剂
ChewBaccaNN: A Flexible 223 TOPS/W BNN Accelerator
摘要：
二进制神经网络使智能物联网设备成为可能，因为它们显著降低了所需的内存占用和计算复杂性，同时保持了高网络性能和灵活性。本文介绍了ChewBaccaNN，一个0.7 mm $^2$大小的二进制卷积神经网络（CNN）加速器，采用GlobalFoundries 22 nm技术设计。通过利用高效的数据重用、数据缓冲、基于锁存器的存储器和电压缩放，实现了241 GOPS的吞吐量，同时在使用高达7 x7内核的二进制CNN进行推理时，在0.4V/154 MHz下仅消耗1.1 mW，从而实现了223 TOPS/W的峰值核心能效。ChewBaccaNN的灵活性允许运行比其他加速器更广泛的二进制CNN，大大提高了TOPS/W度量所能捕获的精度-能量权衡。事实上，它可以以86.8%的准确度执行CIFAR-10推理，仅需1.3 $μJ$，因此超过了准确度，同时与最高效和更大的内存模拟处理设备相比，能源成本降低了2.8倍，同时保持了在需要时运行更大CNN以获得更高准确度的灵活性。它还运行在1000类ILSVRC数据集上训练的二进制ResNet-18，并将能源效率提高了4.4倍。此外，它还可以对使用8基Group-Net训练的二进制化ResNet-18进行推理，以仅3.0 mJ/帧的速度实现67.5%的Top-1准确度-与全精度ResNet-18相比，准确度仅下降1.8%。
英文摘要：
Binary Neural Networks enable smart IoT devices, as they significantly reduce the required memory footprint and computational complexity while retaining a high network performance and flexibility. This paper presents ChewBaccaNN, a 0.7 mm$^2$ sized binary convolutional neural network (CNN) accelerator designed in GlobalFoundries 22 nm technology. By exploiting efficient data re-use, data buffering, latch-based memories, and voltage scaling, a throughput of 241 GOPS is achieved while consuming just 1.1 mW at 0.4V/154MHz during inference of binary CNNs with up to 7x7 kernels, leading to a peak core energy efficiency of 223 TOPS/W. ChewBaccaNN's flexibility allows to run a much wider range of binary CNNs than other accelerators, drastically improving the accuracy-energy trade-off beyond what can be captured by the TOPS/W metric. In fact, it can perform CIFAR-10 inference at 86.8% accuracy with merely 1.3 $μJ$, thus exceeding the accuracy while at the same time lowering the energy cost by 2.8x compared to even the most efficient and much larger analog processing-in-memory devices, while keeping the flexibility of running larger CNNs for higher accuracy when needed. It also runs a binary ResNet-18 trained on the 1000-class ILSVRC dataset and improves the energy efficiency by 4.4x over accelerators of similar flexibility. Furthermore, it can perform inference on a binarized ResNet-18 trained with 8-bases Group-Net to achieve a 67.5% Top-1 accuracy with only 3.0 mJ/frame -- at an accuracy drop of merely 1.8% from the full-precision ResNet-18.
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。
