CADC：面向交叉bar的树突卷积用于高效的内存计算
CADC: Crossbar-Aware Dendritic Convolution for Efficient In-memory Computing
AI总结：
CADC通过在交叉bar计算中嵌入非线性树突函数，显著提高psums稀疏性，减少计算开销并提升能效。
摘要：
卷积神经网络（CNNs）计算密集，通常通过基于交叉bar的内存计算（IMC）架构进行加速。然而，大卷积层必须跨多个交叉bar进行分区，产生大量部分和（psums），需要额外的缓冲、传输和累积，从而引入显著的系统级开销。受神经科学中树突计算原理的启发，我们提出了面向交叉bar的树突卷积（CADC），一种新颖的方法，通过在交叉bar计算中嵌入非线性树突函数（零负值）大幅提高psums的稀疏性。实验结果表明，CADC显著减少了psums，在MNIST上的LeNet-5中减少了80%，在CIFAR-10上的ResNet-18中减少了54%，在CIFAR-100上的VGG-16中减少了66%，在DVS手势数据集上的spiking神经网络（SNN）中减少了高达88%。CADC诱导的稀疏性提供了两个关键优势：（1）使零压缩和零跳过成为可能，从而将缓冲和传输开销减少了29.3%，累积开销减少了47.9%；（2）最小化ADC量化噪声积累，导致小的精度下降——仅LeNet-5为0.01%，ResNet-18为0.1%，VGG-16为0.5%，SNN为0.9%。与传统卷积（vConv）相比，CADC在交叉bar大小从64x64到256x256的范围内，精度变化范围从LeNet-5的+0.11%到+0.19%，ResNet-18的-0.04%到-0.27%，VGG-16的+0.99%到+1.60%，SNN的-0.57%到+1.32%。最终，基于SRAM的IMC实现的CADC在ResNet-18（4/2/4b）上实现了2.15 TOPS和40.8 TOPS/W，相比现有IMC加速器实现了11倍至18倍的速度提升和1.9倍至22.9倍的能量效率提升。
英文摘要：
Convolutional neural networks (CNNs) are computationally intensive and often accelerated using crossbar-based in-memory computing (IMC) architectures. However, large convolutional layers must be partitioned across multiple crossbars, generating numerous partial sums (psums) that require additional buffer, transfer, and accumulation, thus introducing significant system-level overhead. Inspired by dendritic computing principles from neuroscience, we propose crossbar-aware dendritic convolution (CADC), a novel approach that dramatically increases sparsity in psums by embedding a nonlinear dendritic function (zeroing negative values) directly within crossbar computations. Experimental results demonstrate that CADC significantly reduces psums, eliminating 80% in LeNet-5 on MNIST, 54% in ResNet-18 on CIFAR-10, 66% in VGG-16 on CIFAR-100, and up to 88% in spiking neural networks (SNN) on the DVS Gesture dataset. The induced sparsity from CADC provides two key benefits: (1) enabling zero-compression and zero-skipping, thus reducing buffer and transfer overhead by 29.3% and accumulation overhead by 47.9%; (2) minimizing ADC quantization noise accumulation, resulting in small accuracy degradation - only 0.01% for LeNet-5, 0.1% for ResNet-18, 0.5% for VGG-16, and 0.9% for SNN. Compared to vanilla convolution (vConv), CADC exhibits accuracy changes ranging from +0.11% to +0.19% for LeNet-5, -0.04% to -0.27% for ResNet-18, +0.99% to +1.60% for VGG-16, and -0.57% to +1.32% for SNN, across crossbar sizes from 64x64 to 256x256. Ultimately, a SRAM-based IMC implementation of CADC achieves 2.15 TOPS and 40.8 TOPS/W for ResNet-18 (4/2/4b), realizing an 11x-18x speedup and 1.9x-22.9x improvement in energy efficiency compared to existing IMC accelerators.
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。
