{"ID":5675077,"CreatedAt":"2026-07-03T01:40:09.565152011Z","UpdatedAt":"2026-07-05T01:57:11.175896696Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.01607","arxiv_id":"2607.01607","title":"MxGLUT: A Reconfigurable LUT-Centric Broadcast Dataflow Accelerator for Mixed-Precision GEMM","abstract":"Large language model (LLM) inference suffers from growing inefficiency across the prefill and decode phases, especially under weight-only quantization, where activations remain in FP8 while weights are compressed to low-bit integers. Existing LUT-based accelerators mainly target FP8-INT4 computation and still rely on separate floating-point (FP) datapaths for attention GEMM operations, leading to redundant hardware and non-unified mixed-precision execution. Moreover, their static dataflows are poorly matched to the distinct prefill and decode phases. To address these challenges, we propose MxGLUT, a reconfigurable LUT-centric broadcast (RLB) dataflow accelerator built on mixed-precision LUT-based processing elements (MxLPEs). Guided by a unified LUT-based execution framework, MxGLUT organizes both FP8-INT4 and FP8-FP8 GEMMs under a single LUT-based compute mechanism without dedicated FP multipliers or additional FP datapaths, and further adopts the RLB dataflow that localizes heavy partial-sum accumulation during the prefill phase and exploits weight reuse in the decode phase. Synthesized in UMC $28\\,\\mathrm{nm}$ CMOS at $200~\\mathrm{MHz}$, MxGLUT reduces multiplier area by up to $56.92\\%$ and power by up to $77.07\\%$ and $78.35\\%$ in FP8-INT4 and FP8-FP8 modes, respectively. At the accelerator level, MxGLUT achieves an area efficiency of $0.492~\\mathrm{TFLOPS/mm^2}$ and an energy efficiency of $11.58~\\mathrm{TFLOPS/W}$, while adding native FP8-FP8 support incurs only $2.57\\%$ and $3.34\\%$ reductions in area and energy efficiency, respectively, relative to the FP8-INT4-only FIGLUT baseline. Across the Llama family, MxGLUT achieves up to $2.16\\times$ and $1.49\\times$ latency speedup, and reduces normalized energy to $0.44\\times$ and $0.71\\times$ in prefill and decode, respectively, with at most $1.70\\%$ perplexity increase.","short_abstract":"Large language model (LLM) inference suffers from growing inefficiency across the prefill and decode phases, especially under weight-only quantization, where activations remain in FP8 while weights are compressed to low-bit integers. Existing LUT-based accelerators mainly target FP8-INT4 computation and still rely on s...","url_abs":"https://arxiv.org/abs/2607.01607","url_pdf":"https://arxiv.org/pdf/2607.01607v1","authors":"[\"Weiyu Zhou\",\"Chen Ding\",\"Mingyuan Liu\",\"Liangyu Gan\",\"Yukun Feng\",\"Hao Jia\",\"Haoming Chu\",\"Lirong Zheng\",\"Ning Ma\",\"Yuxiang Huan\"]","published":"2026-07-02T02:14:32Z","proceeding":"cs.AR","tasks":"[\"cs.AR\"]","methods":"[\"Large Language Model\",\"Language Model\",\"Generative Adversarial Network\"]","has_code":false}