arXiv Machine Learning
EStream: Fast and Memory-Efficient MoE Prefill through Expert Virtualization on Mobile NPUs
Read the original on arXiv Machine Learning →The Flow has not summarised this story yet — read it at arXiv Machine Learning.
Related stories
arXiv AI
Jul 8
Is Your NPU Ready for LLMs? Dissecting the Hidden Efficiency Bottlenecks in Mobile LLM Inference
arXiv:2607. 05475v1 Announce Type: cross Abstract: Deploying Large Language Models (LLMs) on mobile devices enhances privacy and reduces latency, but is severely bottlenecked by hardware inefficiency.
arXiv AI
Jul 28
DraftExpert: Expansion-Aware Self-Speculative Decoding for End-Device MoE Inference
arXiv:2607. 24434v1 Announce Type: cross Abstract: Large Mixture-of-Experts (MoE) language models are attractive for end-device deployment because only a small subset of experts is active per token, but their routed expert weights often exceed accelerator memory.
arXiv Machine Learning
Jun 11
Energy-Efficient On-Device RAG on a Mobile NPU: System Design and Benchmark on Snapdragon X Elite
arXiv:2606. 11257v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) pipelines are compute-intensive, combining embedding, retrieval, reranking, and large language model (LLM) generation.
arXiv AI
Jul 7
Prima.cpp: Fast 30-70B LLM Inference on Heterogeneous and Low-Resource Home Clusters
arXiv:2504. 08791v3 Announce Type: replace-cross Abstract: On-device inference offers privacy, offline use, and instant response, but consumer hardware restricts large language models (LLMs) to low throughput and capability.
arXiv Machine Learning
Sep 30
Mira: Memory-Efficient MoE Inference Using Adaptive Caching and Predictive Expert Staging
arXiv:2609.38090v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models are a compelling architecture for scaling model capacity, making them especially attractive for deployment on resource-...
arXiv Machine Learning
Sep 25
Paging the Experts: A Reproducible Characterization of Flash-Backed MoE Inference on iPhone
The paper introduces Routide, a Swift/MLX runtime that runs a quantized Qwen3.6-35B-A3B model on iPhone by keeping expert weights on device storage and a byte‑budgeted subset in memory. It evaluates cache‑policy effects, showing that a 512 MiB LRU cache yields 0.00% demand hits while a 576 MiB LRU reaches 38.58% hits across five 128‑token workloads, indicating that capacity limits depend on policy and workload. The study also reports memory footprints, thermal events, and power estimates, demonstrating that flash‑backed MoE inference is feasible within bounded resources but has measurable limitations.