Cover Image

When NPUs Are Not Always Faster: A Stage-Level Analysis of Mobile LLM Inference

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Li, Pu, Qi, Jiawen, Chen, Qinyu
Format:	Preprint
Published:	2026
Subjects:	Hardware Architecture Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

NVR: Vector Runahead on NPUs for Sparse Memory Access
by: Wang, Hui, et al.
Published: (2025)

CaMDN: Enhancing Cache Efficiency for Multi-tenant DNNs on Integrated NPUs
by: Cai, Tianhao, et al.
Published: (2025)

AHASD: Asynchronous Heterogeneous Architecture for LLM Adaptive Drafting Speculative Decoding on Mobile Devices
by: Zirui, Ma, et al.
Published: (2026)

Pushing the Limits of BFP on Narrow Precision LLM Inference
by: Wang, Hui, et al.
Published: (2025)

Comparative Characterization of KV Cache Management Strategies for LLM Inference
by: Mamo, Oteo, et al.
Published: (2026)

FlightLLM: Efficient Large Language Model Inference with a Complete Mapping Flow on FPGAs
by: Zeng, Shulin, et al.
Published: (2024)

T-MAN: Enabling End-to-End Low-Bit LLM Inference on NPUs via Unified Table Lookup
by: Wei, Jianyu, et al.
Published: (2025)

Rethinking LLM Inference Bottlenecks: Insights from Latent Attention and Mixture-of-Experts
by: Yun, Sungmin, et al.
Published: (2025)

Sangam: Chiplet-Based DRAM-PIM Accelerator with CXL Integration for LLM Inferencing
by: Kiyawat, Khyati, et al.
Published: (2025)

Production-Grade Local LLM Inference on Apple Silicon: A Comparative Study of MLX, MLC-LLM, Ollama, llama.cpp, and PyTorch MPS
by: Rajesh, Varun, et al.
Published: (2025)

LUT-LLM: Efficient Large Language Model Inference with Memory-based Computations on FPGAs
by: He, Zifan, et al.
Published: (2025)

PIMphony: Overcoming Bandwidth and Capacity Inefficiency in PIM-based Long-Context LLM Inference System
by: Kwon, Hyucksung, et al.
Published: (2024)

HALO: Memory-Centric Heterogeneous Accelerator with 2.5D Integration for Low-Batch LLM Inference
by: Negi, Shubham, et al.
Published: (2025)

ForgeHLS: A Large-Scale, Open-Source Dataset for High-Level Synthesis
by: Peng, Zedong, et al.
Published: (2025)

Analysis of LLM Vulnerability to GPU Soft Errors: An Instruction-Level Fault Injection Study
by: Chai, Duo, et al.
Published: (2025)

AssertLLM2: A Comprehensive LLM Benchmark for Assertion Generation from Design Specifications
by: Wu, Yuchao, et al.
Published: (2026)

Scalable Processing-Near-Memory for 1M-Token LLM Inference: CXL-Enabled KV-Cache Management Beyond GPU Limits
by: Kim, Dowon, et al.
Published: (2025)

LUMINA: LLM-Guided GPU Architecture Exploration via Bottleneck Analysis
by: Zhang, Tao, et al.
Published: (2026)

Towards LLM-based Root Cause Analysis of Hardware Design Failures
by: Qiu, Siyu, et al.
Published: (2025)

DRC-Coder: Automated DRC Checker Code Generation Using LLM Autonomous Agent
by: Chang, Chen-Chia, et al.
Published: (2024)

FluxEDA: A Unified Execution Infrastructure for Stateful Agentic EDA
by: Chen, Zhengrui, et al.
Published: (2026)

MemExplorer: Navigating the Heterogeneous Memory Design Space for Agentic Inference NPUs
by: Wu, Haoran, et al.
Published: (2026)

Mixture of Cache-Conditional Experts for Efficient Mobile Device Inference
by: Skliar, Andrii, et al.
Published: (2024)

Leveraging Compute-in-Memory for Efficient Generative Model Inference in TPUs
by: Zhu, Zhantong, et al.
Published: (2025)

Silicon Showdown: Performance, Efficiency, and Ecosystem Barriers in Consumer-Grade LLM Inference
by: Javat, Abdurrahman, et al.
Published: (2026)

ELSA: An ELastic SNN Inference Architecture for Efficient Neuromorphic Computing
by: You, Kang, et al.
Published: (2026)

FVDebug: An LLM-Driven Debugging Assistant for Automated Root Cause Analysis of Formal Verification Failures
by: Bai, Yunsheng, et al.
Published: (2025)

iDSE: Navigating Design Space Exploration in High-Level Synthesis Using LLMs
by: Li, Runkai, et al.
Published: (2025)

Accelerating LLM Inference via Dynamic KV Cache Placement in Heterogeneous Memory System
by: Fang, Yunhua, et al.
Published: (2025)

Hybrid Systolic Array Accelerator with Optimized Dataflow for Edge Large Language Model Inference
by: Chen, Chun-Ting, et al.
Published: (2025)

ArchPower: Dataset for Architecture-Level Power Modeling of Modern CPU Design
by: Zhang, Qijun, et al.
Published: (2025)

TL-nvSRAM-CIM: Ultra-High-Density Three-Level ReRAM-Assisted Computing-in-nvSRAM with DC-Power Free Restore and Ternary MAC Operations
by: Wang, Dengfeng, et al.
Published: (2023)

NLS: Natural-Level Synthesis for Hardware Implementation Through GenAI
by: Yang, Kaiyuan, et al.
Published: (2025)

PrefixLLM: LLM-aided Prefix Circuit Design
by: Xiao, Weihua, et al.
Published: (2024)

LLM-DSE: Searching Accelerator Parameters with LLM Agents
by: Wang, Hanyu, et al.
Published: (2025)

Are LLMs Any Good for High-Level Synthesis?
by: Liao, Yuchao, et al.
Published: (2024)

SHIELD: A Segmented Hierarchical Memory Architecture for Energy-Efficient LLM Inference on Edge NPUs
by: Zhang, Jintao, et al.
Published: (2026)

HLS-Eval: A Benchmark and Framework for Evaluating LLMs on High-Level Synthesis Design Tasks
by: Abi-Karam, Stefan, et al.
Published: (2025)

A Two Level Neural Approach Combining Off-Chip Prediction with Adaptive Prefetch Filtering
by: Jamet, Alexandre Valentin, et al.
Published: (2024)

LLM Inference Acceleration via Efficient Operation Fusion
by: Salmani, Mahsa, et al.
Published: (2025)