Agent.xpu: Efficient Scheduling of Agentic LLM Workloads on Heterogeneous SoC

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wei, Xinming, Zhang, Jiahao, Li, Haoran, Chen, Jiayu, Guan, Haoning, Qu, Rui, Li, Maoliang, Chen, Xiang, Luo, Guojie
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911356685385728
author Wei, Xinming
Zhang, Jiahao
Li, Haoran
Chen, Jiayu
Guan, Haoning
Qu, Rui
Li, Maoliang
Chen, Xiang
Luo, Guojie
author_facet Wei, Xinming
Zhang, Jiahao
Li, Haoran
Chen, Jiayu
Guan, Haoning
Qu, Rui
Li, Maoliang
Chen, Xiang
Luo, Guojie
contents Personal LLM agents increasingly combine foreground reactive interactions with background proactive monitoring, forming long-lived, stateful LLM flows that interleave prefill and token-by-token decode. While modern heterogeneous SoCs integrate CPUs, iGPUs, and NPUs to support on-device intelligence, existing LLM engines assume static, single-shot inference and lack mechanisms for flow-level concurrency, prioritization, and efficient accelerator coordination. As a result, commodity SoCs remain poorly matched to the dynamic, mixed-criticality execution patterns of personal agents. This paper presents Agent$.$xpu, the first LLM engine that orchestrates concurrent reactive and proactive LLM flows on commodity SoCs. Extensive profiling uncovers unique SoC characteristics of operator-accelerator affinity, asymmetric DDR contention, and stage-divergent batching behaviors distinct from cloud-serving assumptions. Agent$.$xpu introduces three key techniques: a heterogeneous execution graph (HEG) capturing NPU/iGPU affinity and elastic operator binding; flow-aware NPU-iGPU coordination with stage elasticity, decoupling prefill and decode to reduce bandwidth contention and enforce priorities; and fine-grained preemption with slack-aware piggybacking to guarantee reactive responsiveness without starving proactive work. Across realistic personal-agent workloads, Agent$.$xpu delivers 1.2-4.9$\times$ proactive throughput and reduces reactive latency by at least 91%, compared with both industrial iGPU-only serving engine and NPU-iGPU static inference with optimal tensor-partitioning schemes. Agent$.$xpu also minimizes energy consumption and graphics interference via controlled iGPU usage.
format Preprint
id arxiv_https___arxiv_org_abs_2506_24045
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Agent.xpu: Efficient Scheduling of Agentic LLM Workloads on Heterogeneous SoC
Wei, Xinming
Zhang, Jiahao
Li, Haoran
Chen, Jiayu
Guan, Haoning
Qu, Rui
Li, Maoliang
Chen, Xiang
Luo, Guojie
Distributed, Parallel, and Cluster Computing
Machine Learning
Personal LLM agents increasingly combine foreground reactive interactions with background proactive monitoring, forming long-lived, stateful LLM flows that interleave prefill and token-by-token decode. While modern heterogeneous SoCs integrate CPUs, iGPUs, and NPUs to support on-device intelligence, existing LLM engines assume static, single-shot inference and lack mechanisms for flow-level concurrency, prioritization, and efficient accelerator coordination. As a result, commodity SoCs remain poorly matched to the dynamic, mixed-criticality execution patterns of personal agents. This paper presents Agent$.$xpu, the first LLM engine that orchestrates concurrent reactive and proactive LLM flows on commodity SoCs. Extensive profiling uncovers unique SoC characteristics of operator-accelerator affinity, asymmetric DDR contention, and stage-divergent batching behaviors distinct from cloud-serving assumptions. Agent$.$xpu introduces three key techniques: a heterogeneous execution graph (HEG) capturing NPU/iGPU affinity and elastic operator binding; flow-aware NPU-iGPU coordination with stage elasticity, decoupling prefill and decode to reduce bandwidth contention and enforce priorities; and fine-grained preemption with slack-aware piggybacking to guarantee reactive responsiveness without starving proactive work. Across realistic personal-agent workloads, Agent$.$xpu delivers 1.2-4.9$\times$ proactive throughput and reduces reactive latency by at least 91%, compared with both industrial iGPU-only serving engine and NPU-iGPU static inference with optimal tensor-partitioning schemes. Agent$.$xpu also minimizes energy consumption and graphics interference via controlled iGPU usage.
title Agent.xpu: Efficient Scheduling of Agentic LLM Workloads on Heterogeneous SoC
topic Distributed, Parallel, and Cluster Computing
Machine Learning
url https://arxiv.org/abs/2506.24045