A compact, accurate, and bitwidth-programmable in-memory computing (IMC) static random-access memory (SRAM) macro, named CAP-RAM, is presented for energy-efficient convolutional neural network (CNN) inference. It leverages a novel charge-domain multiply-and-accumulate (MAC) mechanism and circuitry to achieve superior linearity under process variations compared to conventional IMC designs. The adopted semi-parallel architecture efficiently stores filters from multiple CNN layers by sharing eight standard 6T SRAM cells with one charge-domain MAC circuit. Moreover, up to six levels of bit-width of weights with two encoding schemes and eight levels of input activations are supported. A 7-bit charge-injection SAR (ciSAR) analog-to-digital converter (ADC) getting rid of sample and hold (S&H) and input/reference buffers further improves the overall energy efficiency and throughput. A 65-nm prototype validates the excellent linearity and computing accuracy of CAP-RAM. A single $512\times 128$ macro stores a complete pruned and quantized CNN model to achieve 98.8% inference accuracy on the MNIST data set and 89.0% on the CIFAR-10 data set, with a 573.4-giga operations per second (GOPS) peak throughput and a 49.4-tera operations per second (TOPS)/W energy efficiency.

中文翻译：

---

CAP-RAM：用于精确可编程 CNN 推理的电荷域内存计算 6T-SRAM


提出了一种紧凑、精确、位宽可编程内存计算 (IMC) 静态随机存取存储器 (SRAM) 宏，称为 CAP-RAM，用于节能卷积神经网络 (CNN) 推理。与传统 IMC 设计相比，它利用新颖的电荷域乘法累加 (MAC) 机制和电路在工艺变化下实现卓越的线性度。采用的半并行架构通过与一个电荷域 MAC 电路共享 8 个标准 6T SRAM 单元，有效地存储来自多个 CNN 层的滤波器。此外，还支持具有两种编码方案的最多六级权重位宽和八级输入激活。 7 位电荷注入 SAR (ciSAR) 模数转换器 (ADC) 摆脱了采样保持 (S&H) 和输入/参考缓冲器，进一步提高了整体能效和吞吐量。 65 nm 原型验证了 CAP-RAM 出色的线性度和计算精度。单个 $512\times 128$ 宏存储完整的修剪和量化的 CNN 模型，在 MNIST 数据集上实现 98.8% 的推理准确度，在 CIFAR-10 数据集上实现 89.0% 的推理准确度，每秒 573.4 GB 操作 (GOPS) 峰值吞吐量和每秒 49.4 兆次操作 (TOPS)/W 的能效。