TOM: A Ternary Read-only Memory Accelerator for LLM-powered Edge Intelligence

TL;DR

TOM is a novel hybrid ROM-SRAM accelerator leveraging ternary quantization to enable high-density, energy-efficient, real-time LLM inference on edge devices, overcoming memory and bandwidth limitations.

Contribution

It introduces a ternary quantization-based hybrid ROM-SRAM architecture with a sparsity-aware ROM design and workload-aware power management for edge LLM deployment.

Findings

Achieves 3,306 TPS inference throughput with BitNet-2B model.

Demonstrates significant energy efficiency improvements.

Balances model density and tunability for edge AI applications.

Abstract

The deployment of Large Language Models (LLMs) for real-time intelligence on edge devices is rapidly growing. However, conventional hardware architectures face a fundamental memory wall challenge, where limited on-device memory capacity and bandwidth severely constrain the size of deployable models and their inference speed, while also limiting on-device adaptation. To address this challenge, we propose TOM, a hybrid ROM-SRAM accelerator co-designed with ternary quantization, which balances extreme density with on-device tunability. TOM exploits the synergy between ternary quantization and ROM to achieve extreme memory density and bandwidth, while preserving flexibility through a hybrid ROM-SRAM architecture designed for QLoRA-based tunability. Specifically, we introduce: (1) a sparsity-aware ROM architecture that synthesizes ternary weights as standard-cell logic, eliminating area overhead from zero-valued bits; (2) a distributed processing architecture that co-locates high-density ROM banks with flexible SRAM-based QLoRA adapters and compute units; and (3) a workload-aware dynamic power gating scheme that exploits the logic-based nature of ROM to power down inactive banks, minimizing dynamic energy consumption. TOM achieves an inference throughput of 3,306 TPS using BitNet-2B model, demonstrating its effectiveness in delivering real-time, energy-efficient edge intelligence.