Collective I/O Tuning Using Analytical and Machine-Learning Models
Abstract
The ever larger demand of scientific applications for computation and data is currently driving a continuous increase in scale of parallel computers. The inherent complexity of scaling up a computing systems in terms of both hardware and software stack exposes an increasing number of factors impacting the performance and complicating the process of optimization. In particular, the optimization of parallel I/O has become increasingly challenging due to increasing storage hierarchy and well known performance variability of shared storage systems. This paper focuses on model-based autotuning of the two-phase collective I/O algorithm from a popular MPI distribution on the Blue Gene/Q architecture. We propose a novel hybrid model, constructed as a composition of analytical models for communication and storage operations and black-box models for the performance of the individual operations. We perform an in-depth study of the complexity involved in performance modeling including architecture, software stack and noise. In particular we address this challenges of modeling the performance of shared storage systems by building a benchmark that helps synthesizing factors such as topology, file caching, and noise. The experimental results show that the hybrid approach produces significantly better results than state-of-the-art machine learning approaches and shows a higher robustness to noise,more »
- Authors:
- Publication Date:
- Research Org.:
- Argonne National Lab. (ANL), Argonne, IL (United States)
- Sponsoring Org.:
- USDOE Office of Science - Office of Advanced Scientific Computing Research
- OSTI Identifier:
- 1351298
- DOE Contract Number:
- AC02-06CH11357
- Resource Type:
- Conference
- Resource Relation:
- Conference: 2015 IEEE Cluster , 09/08/15 - 09/11/15, Chicago, IL, US
- Country of Publication:
- United States
- Language:
- English
- Subject:
- I/O performance modeling; model-based tuning; statistical and analytical performance models
Citation Formats
Isaila, Florin, Balaprakash, Prasanna, Wild, Stefan M., Kimpe, Dries, Latham, Rob, Ross, Rob, and Hovland, Paul. Collective I/O Tuning Using Analytical and Machine-Learning Models. United States: N. p., 2015.
Web. doi:10.1109/CLUSTER.2015.29.
Isaila, Florin, Balaprakash, Prasanna, Wild, Stefan M., Kimpe, Dries, Latham, Rob, Ross, Rob, & Hovland, Paul. Collective I/O Tuning Using Analytical and Machine-Learning Models. United States. https://doi.org/10.1109/CLUSTER.2015.29
Isaila, Florin, Balaprakash, Prasanna, Wild, Stefan M., Kimpe, Dries, Latham, Rob, Ross, Rob, and Hovland, Paul. 2015.
"Collective I/O Tuning Using Analytical and Machine-Learning Models". United States. https://doi.org/10.1109/CLUSTER.2015.29.
@article{osti_1351298,
title = {Collective I/O Tuning Using Analytical and Machine-Learning Models},
author = {Isaila, Florin and Balaprakash, Prasanna and Wild, Stefan M. and Kimpe, Dries and Latham, Rob and Ross, Rob and Hovland, Paul},
abstractNote = {The ever larger demand of scientific applications for computation and data is currently driving a continuous increase in scale of parallel computers. The inherent complexity of scaling up a computing systems in terms of both hardware and software stack exposes an increasing number of factors impacting the performance and complicating the process of optimization. In particular, the optimization of parallel I/O has become increasingly challenging due to increasing storage hierarchy and well known performance variability of shared storage systems. This paper focuses on model-based autotuning of the two-phase collective I/O algorithm from a popular MPI distribution on the Blue Gene/Q architecture. We propose a novel hybrid model, constructed as a composition of analytical models for communication and storage operations and black-box models for the performance of the individual operations. We perform an in-depth study of the complexity involved in performance modeling including architecture, software stack and noise. In particular we address this challenges of modeling the performance of shared storage systems by building a benchmark that helps synthesizing factors such as topology, file caching, and noise. The experimental results show that the hybrid approach produces significantly better results than state-of-the-art machine learning approaches and shows a higher robustness to noise, at the cost of a higher modeling complexity},
doi = {10.1109/CLUSTER.2015.29},
url = {https://www.osti.gov/biblio/1351298},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Thu Jan 01 00:00:00 EST 2015},
month = {Thu Jan 01 00:00:00 EST 2015}
}