Studying the Degradation of Propagation Delay on FPGAs at the European XFEL

TL;DR

This study empirically analyzes how propagation delay in FPGA devices degrades under harsh conditions at the European XFEL, revealing correlations with radiation exposure and demonstrating machine learning-based degradation estimation.

Contribution

It provides the first empirical analysis of FPGA propagation delay degradation in a high-radiation environment and introduces machine learning models for degradation estimation.

Findings

Degraded FPGA devices exhibit slower switching frequencies.

Radiation doses correlate with increased hardware degradation.

Machine learning models can estimate device switching frequencies.

Abstract

An increasing number of unhardened commercial-off-the-shelf embedded devices are deployed under harsh operating conditions and in highly-dependable systems. Due to the mechanisms of hardware degradation that affect these devices, ageing detection and monitoring are crucial to prevent critical failures. In this paper, we empirically study the propagation delay of 298 naturally-aged FPGA devices that are deployed in the European XFEL particle accelerator. Based on in-field measurements, we find that operational devices show significantly slower switching frequencies than unused chips, and that increased gamma and neutron radiation doses correlate with increased hardware degradation. Furthermore, we demonstrate the feasibility of developing machine learning models that estimate the switching frequencies of the devices based on historical and environmental data.