Skip to yearly menu bar Skip to main content


LUT-GEMM: Quantized Matrix Multiplication based on LUTs for Efficient Inference in Large-Scale Generative Language Models

Gunho Park · baeseong park · Minsub Kim · Sungjae Lee · Jeonghoon Kim · Beomseok Kwon · Se Jung Kwon · Byeongwook Kim · Youngjoo Lee · Dongsoo Lee

Halle B #259
[ ]
Thu 9 May 7:30 a.m. PDT — 9:30 a.m. PDT


Recent advances in self-supervised learning and the Transformer architecture have significantly improved natural language processing (NLP), achieving remarkably low perplexity.However, the growing size of NLP models introduces a memory wall problem during the generation phase.To mitigate this issue, recent efforts have focused on quantizing model weights to sub-4-bit precision while preserving full precision for activations, resulting in practical speed-ups during inference on a single GPU.However, these improvements primarily stem from reduced memory movement, which necessitates a resource-intensive dequantization process rather than actual computational reduction.In this paper, we introduce LUT-GEMM, an efficient kernel for quantized matrix multiplication, which not only eliminates the resource-intensive dequantization process but also reduces computational costs compared to previous kernels for weight-only quantization.Furthermore, we proposed group-wise quantization to offer a flexible trade-off between compression ratio and accuracy.The impact of LUT-GEMM is facilitated by implementing high compression ratios through low-bit quantization and efficient LUT-based operations.We show experimentally that when applied to the OPT-175B model with 3-bit quantization, LUT-GEMM substantially accelerates token generation latency, achieving a remarkable 2.1x improvement on a single GPU when compared to OPTQ, which relies on the costly dequantization process.

Chat is not available.