Timing Predictability in High-Performance Computing With Probabilistic Real-Time

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.

Beyond the Traditional Analyses and Resource Management in Real-Time Systems

“…However, the EVT hypotheses must be satisfied. First preliminary results [13] showed that the EVT hypotheses could be satisfied with due safety technical shrewdness. The presence of heterogeneous hardware (such as GPGPU computing), which is exploding in HPC in the last years, exhibited an improvement in compliance of EVT hypotheses compared to a full-homogeneous scenario.…”

“…What is the relation with embedded safety-critical systems we are talking about in this chapter? Because time-critical applications are also emerging on these systems [13], such as medical imaging, natural disaster prediction, or structures monitoring. All of these application categories require a large amount of computational power (consequently the need to run on HPC clusters) and timing guarantees on the results.…”

RISC-V Processor Technologies for Aerospace Applications in the ISOLDE Project

“…The problem of estimating a safe and tight Worst-Case Execution Time (WCET) in modern processors is an open problem and key issue of real-time embedded systems, and, for certain applications, of HPC computing nodes [26]. The increasing complexity of modern computing platforms (many cores, advanced pipelines, multi-level caches, heterogeneity, etc.)…”

Towards EXtreme scale technologies and accelerators for euROhpc hw/Sw supercomputing applications for exascale: The TEXTAROSSA approach

“…In addition, by integrating the programming model with the resource management layers we could dynamically tune the numerical accuracy (precision) of the algorithms' implementation, with respect to the actual application requirements and power/energy constraints [62]. Similarly, this applies to the problem of guarantee real-time requirements to time-critical applications [61,63]. In general, given the reference hardware platform and the application use-cases, the TEXTAROSSA project would represent an extremely interesting testbench for exploring novel power and energy management strategies, by operating at both software and hardware level.…”