Abstract

The execution of complex distributed applications in exascale systems faces many challenges, as it involves empirical evaluation of countless code variations and application runtime parameters over a heterogeneous set of resources. To mitigate these challenges, the research field of autotuning has gained momentum. The autotuning automates identifying the most desirable application implementation in terms of code variations and runtime parameters. However, the complexity and size of the exascale systems make the autotuning process very difficult, especially considering the number of parameter variations that have to be identified. Therefore, we introduce a novel approach for autotuning exascale applications based on a genetic multi-objective optimization algorithm integrated within the ASPIDE exascale computing framework. The approach considers multi-dimensional search space with support for pluggable objective functions, including execution time and energy requirements. Furthermore, the autotuner employs a machine learning-based event detection approach to detect events and anomalies during application execution, such as hardware failures or communication bottlenecks.

Highlights

  • In the era of the Internet of Things (IoT), massive amounts of data are created continuously, which have to be transferred, processed, and analyzed within tight deadlines

  • As we observe some fastest and slowest solutions with small euclidean distance side by side in the search space, we conclude that the fine-tuning of application parameters for our case study application is highly sensitive for even small parameter value changes and can lead to improvements in the execution time by up to 20%, compared to the standard runtime parameters

  • We introduced an exascale autotuning approach based on an NSGA-II multi-objective optimization algorithm integrated within the ASPIDE exascale computing framework

Read more

Summary

INTRODUCTION

In the era of the Internet of Things (IoT), massive amounts of data are created continuously, which have to be transferred, processed, and analyzed within tight deadlines. The complexity and size of the exascale systems make the autotuning process for exascale applications very difficult, especially considering the number of parameter variations that have to be identified (Durillo and Fahringer, 2014). This problem is further aggravated by the high probability of failures in systems containing thousands of nodes. To address the problem of autotuning, the concurrent approaches are primarily focused on improving the execution time of the application and reducing energy consumption during execution These approaches are not suitable for the exascale systems, as they do not consider the high heterogeneity of the exascale environments.

RELATED WORK
Single-Objective Approaches
Multi-Objective Approaches
Research Gap
Autotuning Process
Implementation
Events and Anomalies Definition
Events and Anomalies Detection Engine Architecture
Anomalies Induction and Detection Methods
ASPIDE SYSTEM ARCHITECTURE
System Interaction
APPLICATION CASE STUDY
Trajectory mining
Autotuner
Pluggable Optimization Objectives
Events and Anomalies Detection Engine
DATA AVAILABILITY STATEMENT
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call