GEREM: Fast and Precise Error Resilience Assessment for GPU Microarchitectures

Jingweijia Tan; Xurui Li; An Zhong; Kaige Yan; Xiaohui Wei; Guanpeng Li

doi:10.1109/TPDS.2025.3552679

Back

GEREM: Fast and Precise Error Resilience Assessment for GPU Microarchitectures

Journal article

Peer reviewed

GEREM: Fast and Precise Error Resilience Assessment for GPU Microarchitectures

Jingweijia Tan, Xurui Li, An Zhong, Kaige Yan, Xiaohui Wei and Guanpeng Li

IEEE transactions on parallel and distributed systems, Vol.36(5), pp.1011-1024

05/2025

DOI: 10.1109/TPDS.2025.3552679

View Online

Abstract

GPUs are widely used hardware acceleration platforms in many areas due to their great computational throughput. In the meanwhile, GPUs are vulnerable to transient hardware faults in the post-Moore era. Analyzing the error resilience of GPUs are critical for both hardware and software. Statistical fault injection approaches are commonly used for error resilience analysis, which are highly accurate but very time consuming. In this work, we propose GEREM, a first framework to speed up fault injection process so as to estimate the error resilience of GPU microarchitectures swiftly and precisely. We find early fault behaviors can be used to accurately predict the final outcomes of program execution. Based on this observation, we categorize the early behaviors of hardware faults into GPU Early Fault Manifestation models (EFMs). For data structures, EFMs are early propagation characteristics of faults, while for pipeline instructions, EFMs are heuristic properties of several instruction contexts. We further observe that EFMs are determined by static microarchitecture states, so we can capture them without actually simulating the program execution process under fault injections. Leveraging these observations, our GEREM framework first profiles the microarchitectural states related for EFMs at one time. It then injects faults into the profiled traces to immediately generate EFMs. For data storage structures, EFMs are directly used to predict final fault outcomes, while for pipeline instructions, machine learning is used for prediction. Evaluation results show GEREM precisely assesses the error resilience of GPU microarchitecture structures with 237\times speedup on average comparing with traditional fault injections.

Software

Accuracy

Benchmark testing

error resilience

fault injection

GPU microarchitecture

Graphics processing units

Hardware

Microarchitecture

Pipelines

Registers

Resilience

Transient analysis

Details

Title: Subtitle: GEREM: Fast and Precise Error Resilience Assessment for GPU Microarchitectures
Creators: Jingweijia Tan - Jilin Province Science and Technology Department
Xurui Li - Jilin University
An Zhong - Jilin Province Science and Technology Department
Kaige Yan - Jilin University
Xiaohui Wei - Jilin Province Science and Technology Department
Guanpeng Li - University of Iowa
Resource Type: Journal article
Publication Details: IEEE transactions on parallel and distributed systems, Vol.36(5), pp.1011-1024
DOI: 10.1109/TPDS.2025.3552679
ISSN: 1045-9219
eISSN: 1558-2183
Publisher: IEEE
Number of pages: 14
Grant note: National Natural Science Foundation of China (NSFC): 62372207
This work was supported in part by the National Natural Science Foundation of China (NSFC) under Grant 62372207.
Language: English
Electronic publication date: 03/17/2025
Date published: 05/2025
Academic Unit: Computer Science
Record Identifier: 9984802410202771

Metrics

6 Record Views