CodeGEMM: A Codebook-Centric Approach to Efficient GEMM in Quantized LLMs

Gunho Park, Jeongin Bae, Byeongwook Kim, Baeseong park, Jiwon Ryu, Hoseung Kim, Se Jung Kwon, Dongsoo Lee

Keywords: Quantization, Matrix Multiplication, Large Language Models

Abstract:

Weight-only quantization is widely used to mitigate the memory-bound nature of LLM inference. Codebook-based methods extend this trend by achieving strong accuracy in the extremely low-bit regime (e.g., 2-bit). However, current kernels rely on dequantization, which repeatedly fetches centroids and reconstructs weights, incurring substantial latency and cache pressure. We present CodeGEMM, a codebook-centric GEMM kernel that replaces dequantization with precomputed inner products between centroids and activations stored in a lightweight Psumbook. At inference, code indices directly gather these partial sums, eliminating per-element lookups and reducing the on-chip footprint. The kernel supports the systematic exploration of latency–memory–accuracy trade-offs under a unified implementation.
On Llama-3 models, CodeGEMM delivers 1.83x (8B) and 8.93x (70B) speedups in the 2-bit configuration compared to state-of-the-art codebook-based quantization at comparable accuracy and further improves computing efficiency and memory subsystem utilization.

Primary Area: Deep learning (e.g., architectures, generative models, optimization for deep networks, foundation models, LLMs)

Submission Number: 11257