{"ID":2826927,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.17970","arxiv_id":"2512.17970","title":"CodeGEMM: A Codebook-Centric Approach to Efficient GEMM in Quantized LLMs","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.","short_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,...","url_abs":"https://arxiv.org/abs/2512.17970","url_pdf":"https://arxiv.org/pdf/2512.17970v1","authors":"[\"Gunho Park\",\"Jeongin Bae\",\"Byeongwook Kim\",\"Baeseong park\",\"Jiwon Ryu\",\"Hoseung Kim\",\"Se Jung Kwon\",\"Dongsoo Lee\"]","published":"2025-12-19T06:16:32Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\"]","methods":"[\"Large Language Model\",\"LoRA\"]","has_code":false}