Understanding Data Movement in AMD Multi-GPU Systems with Infinity Fabric

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Schieffer, Gabin, Shi, Ruimin, Markidis, Stefano, Herten, Andreas, Faj, Jennifer, Peng, Ivy
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910627561209856
author Schieffer, Gabin
Shi, Ruimin
Markidis, Stefano
Herten, Andreas
Faj, Jennifer
Peng, Ivy
author_facet Schieffer, Gabin
Shi, Ruimin
Markidis, Stefano
Herten, Andreas
Faj, Jennifer
Peng, Ivy
contents Modern GPU systems are constantly evolving to meet the needs of computing-intensive applications in scientific and machine learning domains. However, there is typically a gap between the hardware capacity and the achievable application performance. This work aims to provide a better understanding of the Infinity Fabric interconnects on AMD GPUs and CPUs. We propose a test and evaluation methodology for characterizing the performance of data movements on multi-GPU systems, stressing different communication options on AMD MI250X GPUs, including point-to-point and collective communication, and memory allocation strategies between GPUs, as well as the host CPU. In a single-node setup with four GPUs, we show that direct peer-to-peer memory accesses between GPUs and utilization of the RCCL library outperform MPI-based solutions in terms of memory/communication latency and bandwidth. Our test and evaluation method serves as a base for validating memory and communication strategies on a system and improving applications on AMD multi-GPU computing systems.
format Preprint
id arxiv_https___arxiv_org_abs_2410_00801
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Understanding Data Movement in AMD Multi-GPU Systems with Infinity Fabric
Schieffer, Gabin
Shi, Ruimin
Markidis, Stefano
Herten, Andreas
Faj, Jennifer
Peng, Ivy
Distributed, Parallel, and Cluster Computing
Modern GPU systems are constantly evolving to meet the needs of computing-intensive applications in scientific and machine learning domains. However, there is typically a gap between the hardware capacity and the achievable application performance. This work aims to provide a better understanding of the Infinity Fabric interconnects on AMD GPUs and CPUs. We propose a test and evaluation methodology for characterizing the performance of data movements on multi-GPU systems, stressing different communication options on AMD MI250X GPUs, including point-to-point and collective communication, and memory allocation strategies between GPUs, as well as the host CPU. In a single-node setup with four GPUs, we show that direct peer-to-peer memory accesses between GPUs and utilization of the RCCL library outperform MPI-based solutions in terms of memory/communication latency and bandwidth. Our test and evaluation method serves as a base for validating memory and communication strategies on a system and improving applications on AMD multi-GPU computing systems.
title Understanding Data Movement in AMD Multi-GPU Systems with Infinity Fabric
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2410.00801