用并行框架分析大规模GPU日志中的性能波动，提升分析效率67%。
A Distributed Framework for Causal Modeling of Performance Variability in GPU Traces
- 将大尺度GPU日志分块并行处理，结合因果图与协调图揭示执行依赖
- 在多日志独立分析场景下实现67%的可扩展性提升
- 适合高性能计算领域做性能瓶颈诊断的研究者与工程师
大规模GPU日志在识别异构高性能计算（HPC）架构中的性能瓶颈中起着关键作用。然而，单个日志数据量庞大且复杂，导致性能分析计算成本高、耗时长。为此，我们提出一个端到端的并行性能分析框架，可高效处理多个大规模GPU日志。该框架对日志数据进行分片并行处理，结合因果图方法与并行协调图，揭示执行流程间的性能波动与依赖关系。实验结果表明，在多日志独立分析场景下，系统可扩展性提升67%，验证了该流水线的有效性。
原文摘要 · Abstract (English)
Large-scale GPU traces play a critical role in identifying performance bottlenecks within heterogeneous High-Performance Computing (HPC) architectures. However, the sheer volume and complexity of a single trace of data make performance analysis both computationally expensive and time-consuming. To address this challenge, we present an end-to-end parallel performance analysis framework designed to handle multiple large-scale GPU traces efficiently. Our proposed framework partitions and processes trace data concurrently and employs causal graph methods and parallel coordinating chart to expose performance variability and dependencies across execution flows. Experimental results demonstrate a 67% improvement in terms of scalability, highlighting the effectiveness of our pipeline for analyzing multiple traces independently.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。