Understanding Error Propagation in GPGPU Applications

Guanpeng Li; Karthik Pattabiraman; Chen-Yang Cher; Pradip Bose

doi:10.1109/SC.2016.20

Back

Conference proceeding

Understanding Error Propagation in GPGPU Applications

Guanpeng Li, Karthik Pattabiraman, Chen-Yang Cher and Pradip Bose

SC16: International Conference for High Performance Computing, Networking, Storage and Analysis, pp.240-251

11/2016

DOI: 10.1109/SC.2016.20

View Online

Abstract

GPUs have emerged as general-purpose accelerators in high-performance computing (HPC) and scientific applications. However, the reliability characteristics of GPU applications have not been investigated in depth. While error propagation has been extensively investigated for non-GPU applications, GPU applications have a very different programming model which can have a significant effect on error propagation in them. We perform an empirical study to understand and characterize error propagation in GPU applications. We build a compilerbased fault-injection tool for GPU applications to track error propagation, and define metrics to characterize propagation in GPU applications. We find GPU applications exhibit significant error propagation for some kinds of errors, but not others, and the behaviour is highly application specific. We observe the GPUCPU interaction boundary naturally limits error propagation in these applications compared to traditional non-GPU applications. We also formulate various guidelines for the design of faulttolerance mechanisms in GPU applications based on our results.

Central Processing Unit

Circuit faults

Computational modeling

CUDA

Error Propagation

Error Resilience

Fault Injection

GPGPU

Graphics processing units

Hardware

Kernel

Programming

Details

Title: Subtitle: Understanding Error Propagation in GPGPU Applications
Creators: Guanpeng Li - University of British Columbia
Karthik Pattabiraman - University of British Columbia
Chen-Yang Cher - T.J. Watson Res. Center, IBM, Yorktown Heights, NY, USA
Pradip Bose - T.J. Watson Res. Center, IBM, Yorktown Heights, NY, USA
Resource Type: Conference proceeding
Publication Details: SC16: International Conference for High Performance Computing, Networking, Storage and Analysis, pp.240-251
Publisher: IEEE
DOI: 10.1109/SC.2016.20
ISSN: 2167-4329
eISSN: 2167-4337
Language: English
Date published: 11/2016
Academic Unit: Computer Science
Record Identifier: 9984259460502771

Metrics

48 Record Views