Weight-only quantization has emerged as a promising solution to the deployment challenges of large language models (LLMs). However, it necessitates FP-INT operations, which make implementation on general-purpose hardware like GPUs difficult. In this paper, we propose FIGLUT, an efficient look-up table (LUT)-based GEMM accelerator architecture. Instead of performing traditional arithmetic operations, FIGLUT retrieves precomputed values from an LUT based on weight patterns, significantly reducing the computational complexity. We also introduce a novel LUT design that addresses the limitations of conventional memory architectures. To further improve LUT-based operations, we propose a half-size LUT combined with a dedicated decoding and multiplexing unit. FIGLUT efficiently supports different bit precisions and quantization methods using a single fixed hardware configuration. For the same 3-bit weight precision, FIGLUT demonstrates 59% higher TOPS/W and 20% lower perplexity than state-of-the-art accelerator designs. When targeting the same perplexity, FIGLUT achieves 98% higher TOPS/W by performing 2.4-bit operations.
@article{arxiv.2503.06862,
title = {FIGLUT: An Energy-Efficient Accelerator Design for FP-INT GEMM Using Look-Up Tables},
author = {Gunho Park and Hyeokjun Kwon and Jiwoo Kim and Jeongin Bae and Baeseong Park and Dongsoo Lee and Youngjoo Lee},
journal= {arXiv preprint arXiv:2503.06862},
year = {2025}
}