ARTEMIS: A Mixed Analog-Stochastic In-DRAM Accelerator for Transformer Neural Networks

TL;DR

ARTEMIS is a novel in-DRAM accelerator that combines analog and stochastic computing to significantly speed up transformer neural networks while reducing energy consumption, addressing the computational and memory challenges of traditional hardware.

Contribution

It introduces a mixed analog-stochastic in-DRAM architecture supporting transformer models with minimal DRAM modifications, enabling efficient in-memory processing.

Findings

Achieves at least 3.0x speedup over GPUs and TPUs

Reduces energy consumption by 1.8x compared to state-of-the-art hardware

Offers 1.9x better energy efficiency than existing accelerators

Abstract

Transformers have emerged as a powerful tool for natural language processing (NLP) and computer vision. Through the attention mechanism, these models have exhibited remarkable performance gains when compared to conventional approaches like recurrent neural networks (RNNs) and convolutional neural networks (CNNs). Nevertheless, transformers typically demand substantial execution time due to their extensive computations and large memory footprint. Processing in-memory (PIM) and near-memory computing (NMC) are promising solutions to accelerating transformers as they offer high compute parallelism and memory bandwidth. However, designing PIM/NMC architectures to support the complex operations and massive amounts of data that need to be moved between layers in transformer neural networks remains a challenge. We propose ARTEMIS, a mixed analog-stochastic in-DRAM accelerator for transformer models. Through employing minimal changes to the conventional DRAM arrays, ARTEMIS efficiently alleviates the costs associated with transformer model execution by supporting stochastic computing for multiplications and temporal analog accumulations using a novel in-DRAM metal-on-metal capacitor. Our analysis indicates that ARTEMIS exhibits at least 3.0x speedup, 1.8x lower energy, and 1.9x better energy efficiency compared to GPU, TPU, CPU, and state-of-the-art PIM transformer hardware accelerators.