NAVIgator under Fire: Characterizing Voltage Limits and Soft Error Resilience of AMD NAVI GPUs
- 26 February 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Device and Materials Reliability
- Vol. PP (15304388) , 1
Abstract
As semiconductor fabrication scales to smaller technology nodes, process variation presents significant challenges, affecting power consumption, thermal behavior, and voltage stability in silicon compute devices. Conservative voltage guardbands are traditionally used to ensure reliable operation under worst-case process, voltage, and temperature (PVT) variations, but they lead to excessive power consumption. Reducing the supply voltage is a promising technique for improving energy efficiency without sacrificing computational correctness and performance. While extensive research has been conducted on reducing the voltage levels in CPUs and NVIDIA GPUs, the AMD GPU side remains relatively unexplored. Process variation, inherent in semiconductor manufacturing, results in differences in power efficiency, thermal characteristics, and voltage stability even among identical GPUs from the same production batch. In this paper, we present an extensive study on voltage scaling beyond nominal conditions for three modern AMD NAVI GPUs (i.e., RX 7600 XT, 7700 XT, and 7800 XT) executing both conventional benchmarks and PyTorch-based machine learning workloads. We evaluate and present power savings and execution stability by combining undervolting characterization with neutron irradiation physical experiments, contributing to a deeper understanding of undervolting and resilience of AMD GPUs.Keywords
Funding Information
- Hellenic Foundation Research and Innovation (HFRI) Project Leveraging Voltage Margins and Error Resiliency at Large-Scale Computing
- Chips Joint Undertaking (JU) Program (101097224 (REBECCA))
- European High-Performance Computing (EuroHPC) JU Program (101202459 (DARE-SGA1))
- Advanced Micro Devices (AMD), Meta, and the Open Compute Project
This publication has 39 references indexed in Scilit:
- ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices2018
- Measuring and Exploiting Guardbands of Server-Grade ARMv8 CPU Cores and DRAMs2018
- A Framework for Evaluating Software on Reduced Margins Hardware2018
- Fault and Error Tolerance in Neural Networks: A ReviewIEEE Access, 2017
- Densely Connected Convolutional Networks2017
- Deep Residual Learning for Image Recognition2016
- Going deeper with convolutions2015
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998
- Threshold-voltage sensitivity of ion-implanted m.o.s. transistors due to process variationsElectronics Letters, 1974