CORTEXA
← Browse
arxivcs.AR2026-07-25

Decoding the Skew: Distribution-Aware MoE Inference with Adaptive Kernel Dispatch

En-Ming Huang, An-Cheng Chang, Bai-Cheng Jeng, Shih-Hao Hung, H. T. Kung

Mixture-of-Experts (MoE) inference consists of sparse expert GEMMs whose shapes vary with the runtime routing distribution. Existing serving systems typically select fused-MoE kernels using static token-count buckets, ignoring the per-expert routing distribution that determines tile padding, memory reuse, and kernel efficiency. We introduce a distribution-aware framework for modeling and benchmarking MoE inference. The framework combines the compact Effective Experts metric with a Dirichlet-based reverse-modeling procedure that generates controllable routing distributions for systematic hardware studies. Using it, we show that the best fused-MoE kernel changes with routing skew and token count. We further present DA-MoE, a GPU-resident kernel-dispatch runtime for NVIDIA GPUs that matches the live routing histogram to offline-tuned distributions and selects a near-optimal fused-MoE kernel without CPU--GPU synchronization. On HumanEval-X serving traces, DA-MoE improves geomean fused-MoE latency by 1.16X on DeepSeek-V3 and 1.29X on Kimi K2, with peak speedups of 1.40X and 1.56X.

View free PDFSource page

Related papers

arxivcs.ARcs.LG2026-07-07

HiFuzz: Hierarchical Reinforcement Learning for Semantic-Aware and Adaptive CPU Fuzzing

Ya Wang, Hanwei Fan, Zhenguo Liu, Xiaofeng Zhou, Yangdi Lyu, Jiang Xu, et al.

Modern processor verification struggles to reach deep architectural states due to the inefficiencies of traditional mutation-based fuzzing. We propose HiFuzz, a novel hierarchical reinforcement learning framework that replaces mutation with a structured, two-layer generation proc…

View free PDFSource page
arxivcs.LGcs.ARcs.CV2026-07-20

Empowering On-Device Model Adaptation with an Edge AI Inference Accelerator

Mateusz Piechocki, Alessandro Capotondi, Marek Kraft

On-device model adaptation is essential to enable lifelong personalization on resource-constrained hardware, but compute, power, and memory limitations of such devices make end-to-end backpropagation impractical for modern deep neural networks. This work proposes a heterogeneous…

View free PDFSource page
arxivcs.PFcs.AIcs.AR2026-06-26

KernelSight-LM: A Kernel-Level LLM Inference Simulator

Xiteng Yao, Taeho Kim, Hengzhi Pei, Xinle Liu, Kyle Ulrich, Leonard Lausen, et al.

As large language models (LLMs) move into production serving, practitioners must rapidly evaluate inference performance across diverse hardware, models, and serving parameters to meet cost and latency targets. However, the end-to-end behavior of LLMs couples serving-layer policie…

View free PDFSource page
arxivcs.ARcs.ETcs.LG2026-07-19

ThRIve: Thermally Robust CNN Inference via Low-Rank Adaptation in Heterogeneous PIM Architectures

Vibhanshu Sharma, Pratyush Dhingra, Janardhan Rao Doppa, Partha Pratim Pande

Processing-In-Memory (PIM) has emerged as a promising technology for accelerating machine learning (ML) workloads. Specifically, non-volatile memory-based PIM architectures have enabled effective ML acceleration due to their ability to perform energy-efficient matrix-vector multi…

View free PDFSource page
arxivcs.AR2026-07-28

Beyond Prefill-Decode Disaggregation: Dissecting LLM Inference for Heterogeneous Platforms via Dynamic Operator Scheduling

Jiaqi Yang, Jiayi Li, Yihan Fu, Hongxiao Zhao, Zhan Chen, Qiuping Wu, et al.

Prefill-decode disaggregation (PD) and roofline-based operator placement are common strategies for partitioning Large Language Model (LLM) inference across heterogeneous systems, but they are often insufficient in practice. End-to-end latency also depends on workload shape, runti…

View free PDFSource page