Abstract
Application requirements in High-Performance Computing (HPC) are becoming increasingly exacting, and the demand for computational resources is rising. In parallel, new application domains are emerging, as well as additional requirements, such as meeting real-time constraints. This requirement, typical of embedded systems, is difficult to guarantee when dealing with HPC infrastructures, due to the intrinsic complexity of the system. Traditional embedded systems static analyses to estimate the Worst-Case Execution Time (WCET) are not applicable to HPC, because modeling and analyzing all the system’s hardware and software components is not practical. Measurement-based probabilistic analyses for the WCET emerged in the last decade to overcome these issues, but it requires the system to satisfy certain conditions to estimate a correct and safe WCET. In this work, we show the emerging application timing requirements, and we propose to exploit the probabilistic real-time theory to achieve the required time predictability. After a brief recap of the fundamentals of this methodology, we focus on its applicability to HPC systems to check their ability to satisfy such conditions. In particular, we studied the advantages of having heterogeneous processors in HPC nodes and how resource management affects the applicability of the proposed technique.
Highlights
High-Performance Computing (HPC) aims at providing computing infrastructures capable of fulfilling the increasing performance requirements of modern applications, in both scientific and industrial domains
THE PROPOSED SOLUTION AND CONTRIBUTIONS To guarantee the timing predictability of the applications, without impacting the HPC infrastructure and its overall average performance, we propose relying on an approach that has come out recently in the real-time Cyber-Physical System (CPS)3 world: Probabilistic Real-Time Computing
2) propose to use the probabilistic real-time theory on the HPC applications instead of the embedded systems, and we introduce the necessary background of this technique for the reader; 3) identify the barriers for the exploitation of probabilistic real-time in HPC systems, and we discuss how heterogeneous computing may be helpful to solve such issues; 4) analyze, both qualitatively and quantitatively, the satisfaction of the probabilistic real-time hypotheses for different resource management choices, basing the evaluation on experiments performed on a real HPC cluster; 5) discuss the different benefits of this technique in improving the meeting of the application requirements, and the management of the job queue and, in general, the HPC resources
Summary
High-Performance Computing (HPC) aims at providing computing infrastructures capable of fulfilling the increasing performance requirements of modern applications, in both scientific and industrial domains. As a consequence of the end of Dennard scaling, maximizing energy efficiency has become the primary goal of the evolution of high-performance processors To achieve this goal, processor manufacturers added many advanced features (e.g., pipelines, multi-level caches, vector instructions). Processor manufacturers added many advanced features (e.g., pipelines, multi-level caches, vector instructions) In this picture, having both a distributed topology and highperformance multi-core processors, HPC systems offer the possibility of scaling the performance of the applications by leveraging on both inter-node and intra-node parallelisms. To effectively exploit such parallelisms, proper software frameworks are required [1].
Sign-in/Register to view highlights & summary 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 callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
