Hemlet: A Heterogeneous Compute-in-Memory Chiplet Architecture for Vision Transformers with Group-Level Parallelism

By Cong Wang, Zexin Fu, Jiayi Huang, and Shanshi Huang
The Hong Kong University of Science and Technology (Guangzhou)

Abstract

Vision Transformers (ViTs) have established new performance benchmarks in vision tasks such as image recognition and object detection. However, these advancements come with significant demands for memory and computational resources, presenting challenges for hardware deployment. Heterogeneous compute-in-memory (CIM) accelerators have emerged as a promising solution for enabling energy-efficient deployment of ViTs. Despite this potential, monolithic CIM-based designs face scalability issues due to the size limitations of a single chip. To address this challenge, emerging chiplet-based techniques offer a more scalable alternative. However, chiplet designs come with their own costs, as they introduce more expensive communication through network-on-package (NoP) compared to network-onchip (NoC), which can hinder improvements in throughput.

This work introduces Hemlet—a heterogeneous CIM chiplet system designed to accelerate ViT. Hemlet facilitates flexible resource scaling through the integration of heterogeneous analog CIM (ACIM), digital CIM (DCIM), and Intermediate Data Process (IDP) chiplets. To improve throughput while reducing communication overhead, it employs a group-level parallelism (GLP) mapping strategy and system-level dataflow optimization, achieving speedups ranging from 1.44× to 4.07× across various hardware configurations within the chiplet system. Our evaluation results demonstrate that Hemlet can achieve a throughput of 8.68 TOPS with an energy efficiency of 3.86 TOPS/W.

Index Terms—Compute-in-memory, Chiplet, Heterogeneous Computing, Mapping

To read the full article, click here