eACGM: Non-instrumented Performance Tracing and Anomaly Detection towards Machine Learning Systems

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Xu, Ruilin, Xie, Zongxuan, Chen, Pengfei
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915367728709632
author Xu, Ruilin
Xie, Zongxuan
Chen, Pengfei
author_facet Xu, Ruilin
Xie, Zongxuan
Chen, Pengfei
contents We present eACGM, a full-stack AI/ML system monitoring framework based on eBPF. eACGM collects real-time performance data from key hardware components, including the GPU and network communication layer, as well as from key software stacks such as CUDA, Python, and PyTorch, all without requiring any code instrumentation or modifications. Additionally, it leverages libnvml to gather process-level GPU resource usage information. By applying a Gaussian Mixture Model (GMM) to the collected multidimensional performance metrics for statistical modeling and clustering analysis, eACGM effectively identifies complex failure modes, such as latency anomalies, hardware failures, and communication inefficiencies, enabling rapid diagnosis of system bottlenecks and abnormal behaviors. To evaluate eACGM's effectiveness and practicality, we conducted extensive empirical studies and case analyses in multi-node distributed training scenarios. The results demonstrate that eACGM, while maintaining a non-intrusive and low-overhead profile, successfully captures critical performance anomalies during model training and inference. Its stable anomaly detection performance and comprehensive monitoring capabilities validate its applicability and scalability in real-world production environments, providing strong support for performance optimization and fault diagnosis in large-scale AI/ML systems.
format Preprint
id arxiv_https___arxiv_org_abs_2506_02007
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle eACGM: Non-instrumented Performance Tracing and Anomaly Detection towards Machine Learning Systems
Xu, Ruilin
Xie, Zongxuan
Chen, Pengfei
Distributed, Parallel, and Cluster Computing
Artificial Intelligence
Networking and Internet Architecture
We present eACGM, a full-stack AI/ML system monitoring framework based on eBPF. eACGM collects real-time performance data from key hardware components, including the GPU and network communication layer, as well as from key software stacks such as CUDA, Python, and PyTorch, all without requiring any code instrumentation or modifications. Additionally, it leverages libnvml to gather process-level GPU resource usage information. By applying a Gaussian Mixture Model (GMM) to the collected multidimensional performance metrics for statistical modeling and clustering analysis, eACGM effectively identifies complex failure modes, such as latency anomalies, hardware failures, and communication inefficiencies, enabling rapid diagnosis of system bottlenecks and abnormal behaviors. To evaluate eACGM's effectiveness and practicality, we conducted extensive empirical studies and case analyses in multi-node distributed training scenarios. The results demonstrate that eACGM, while maintaining a non-intrusive and low-overhead profile, successfully captures critical performance anomalies during model training and inference. Its stable anomaly detection performance and comprehensive monitoring capabilities validate its applicability and scalability in real-world production environments, providing strong support for performance optimization and fault diagnosis in large-scale AI/ML systems.
title eACGM: Non-instrumented Performance Tracing and Anomaly Detection towards Machine Learning Systems
topic Distributed, Parallel, and Cluster Computing
Artificial Intelligence
Networking and Internet Architecture
url https://arxiv.org/abs/2506.02007