ucTrace: A Multi-Layer Profiling Tool for UCX-driven Communication

TL;DR

ucTrace is a new profiling tool that provides detailed, fine-grained visualization of UCX-driven communication in HPC systems, aiding optimization and debugging of large-scale MPI applications.

Contribution

The paper introduces ucTrace, a novel profiler that visualizes UCX communication at a fine-grained level, linking MPI operations to device and transport-layer behaviors.

Findings

ucTrace effectively visualizes MPI communication patterns.

It reveals transport-layer behaviors and device interactions.

Demonstrates utility in optimizing large-scale HPC applications.

Abstract

UCX is a communication framework that enables low-latency, high-bandwidth communication in HPC systems. With its unified API, UCX facilitates efficient data transfers across multi-node CPU-GPU clusters. UCX is widely used as the transport layer for MPI, particularly in GPU-aware implementations. However, existing profiling tools lack fine-grained communication traces at the UCX level, do not capture transport-layer behavior, or are limited to specific MPI implementations. To address these gaps, we introduce ucTrace, a novel profiler that exposes and visualizes UCX-driven communication in HPC environments. ucTrace provides insights into MPI workflows by profiling message passing at the UCX level, linking operations between hosts and devices (e.g., GPUs and NICs) directly to their originating MPI functions. Through interactive visualizations of process- and device-specific interactions, ucTrace helps system administrators, library and application developers optimize performance and debug communication patterns in large-scale workloads. We demonstrate ucTrace's features through a wide range of experiments including MPI point-to-point behavior under different UCX settings, Allreduce comparisons across MPI libraries, communication analysis of a linear solver, NUMA binding effects, and profiling of GROMACS MD simulations with GPU acceleration at scale. ucTrace is publicly available at https://github.com/ParCoreLab/ucTrace.