Exploring The Effect Of Compiler Optimizations On The Reliability Of Hpc Applications
Keywords
compiler; Exascale; faults; HPC; llvm; resilience
Abstract
The strict power efficiency constraints required to achieve exascale systems will dramatically increase the number of detected and undetected transient errors in future high performance computing (HPC) systems. Among the various factors that effect system resiliency, the impact of compiler optimizations on the vulnerability of scientific applications executed on HPC systems has not been widely explored. In this work, we analyze whether and how most common compiler optimizations impact the vulnerability of several mission-critical applications, what are the trade-offs between performance and vulnerability and the causal relations between compiler optimization and application vulnerability. We show that highly-optimized code is generally more vulnerable than unoptimized code. We also show that, while increasing optimization level can drastically improve application performance as expected. However, certain cases of optimization may provide only marginal benefits, but considerably increase application vulnerability.
Publication Date
6-30-2017
Publication Title
Proceedings - 2017 IEEE 31st International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2017
Number of Pages
1274-1283
Document Type
Article; Proceedings Paper
Personal Identifier
scopus
DOI Link
https://doi.org/10.1109/IPDPSW.2017.7
Copyright Status
Unknown
Socpus ID
85028073762 (Scopus)
Source API URL
https://api.elsevier.com/content/abstract/scopus_id/85028073762
STARS Citation
Ashraf, Rizwan A.; Gioiosa, Roberto; Kestor, Gokcen; and DeMara, Ronald F., "Exploring The Effect Of Compiler Optimizations On The Reliability Of Hpc Applications" (2017). Scopus Export 2015-2019. 6960.
https://stars.library.ucf.edu/scopus2015/6960