PCLVis: Visual Analytics of Process Communication Latency in Large-Scale Simulation

2025-06-29

AI Summary

Large-scale parallel simulations on supercomputers are often bottlenecked by inter-process communication latency (PCL), yet existing diagnostic methods require administrator-level access to physical network topology—making them inaccessible to ordinary users. To address this, we propose PCLVis, the first user-oriented visual analytics framework for PCL diagnosis. It automatically constructs a process correlation tree and a communication-dependency DAG from MPI logs, and introduces two novel abstractions: a sliding-window algorithm for temporal segmentation and Communication-State Glyphs (CS-Glyphs) for encoding latency states. These enable spatiotemporal localization of PCL events, propagation-path tracing, and root-cause attribution. Furthermore, clustering and graph-based modeling support latency-pattern recognition and interactive exploration. Evaluated on multiple real-world simulation workloads on the TH-1A supercomputer, PCLVis significantly improves PCL diagnosis efficiency and delivers a deployable, user-accessible tool for performance optimization of large-scale simulations.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Web Mining and Content Analysis: Web data visualization

Abstract

Large-scale simulations on supercomputers have become important tools for users. However, their scalability remains a problem due to the huge communication cost among parallel processes. Most of the existing communication latency analysis methods rely on the physical link layer information, which is only available to administrators. In this paper, a framework called PCLVis is proposed to help general users analyze process communication latency (PCL) events. Instead of the physical link layer information, the PCLVis uses the MPI process communication data for the analysis. First, a spatial PCL event locating method is developed. All processes with high correlation are classified into a single cluster by constructing a process-correlation tree. Second, the propagation path of PCL events is analyzed by constructing a communication-dependency-based directed acyclic graph (DAG), which can help users interactively explore a PCL event from the temporal evolution of a located PCL event cluster. In this graph, a sliding window algorithm is designed to generate the PCL events abstraction. Meanwhile, a new glyph called the communication state glyph (CS-Glyph) is designed for each process to show its communication states, including its in/out messages and load balance. Each leaf node can be further unfolded to view additional information. Third, a PCL event attribution strategy is formulated to help users optimize their simulations. The effectiveness of the PCLVis framework is demonstrated by analyzing the PCL events of several simulations running on the TH-1A supercomputer. By using the proposed framework, users can greatly improve the efficiency of their simulations.

Problem

• Analyzes process communication latency in large-scale simulations
• Identifies high-correlation processes using MPI data
• Visualizes communication paths and states for optimization

Innovation

• Uses MPI process communication data for analysis
• Develops process-correlation tree for event clustering
• Designs communication-dependency-based DAG for path analysis