Introduction to RADR 2019 - Archive ouverte HAL Access content directly
Documents Associated With Scientific Events Year :

Introduction to RADR 2019

(1) , (2) , (1)
1
2

Abstract

The question of efficient dynamic allocation of compute-node resources, such as cores, by independent libraries or runtime systems can be an nightmare. Scientists writing application components have no way to efficiently specify and compose resource-hungry components. As application software stacks become deeper and the interaction of multiple runtime layers compete for resources from the operating system, it has become clear that intelligent cooperation is needed. Resources such as compute cores, in-package memory, and even electrical power must be orchestrated dynamically across application components, with the ability to query each other and respond appropriately. A more integrated solution would reduce intra-application resource competition and improve performance. Furthermore, application runtime systems could request and allocate specific hardware assets and adjust runtime tuning parameters up and down the software stack. The goal of this workshop is to gather and share the latest scholarly research from the community working on these issues, at all levels of the HPC software stack. This include thread allocation, resource arbitration and management, containers, and so on, from runtime-system designers to compilers. We will also use panel sessions and keynote talks to discuss these issues, share visions, and present solutions. Scope Over the last five years, the number of nodes in large supercomputers has remained largely unchanged. In fact, the Oak Ridge National Laboratory computer leading the Top500 list, Summit, has fewer nodes than its predecessor, which is 20 times slower. Machines are getting faster not by adding nodes, but by adding parallelism, cores, and hierarchical memory to each compute node. This shift in how computers are scaled up makes it imperative that parallel computer resources within a node be carefully orchestrated to achieve maximum performance. Dynamically allocating and managing threads and the mapping of these threads to cores is a challenge that requires cooperation and coordination between the different components of the software stack.
Fichier principal
Vignette du fichier
RADR Intro2019.pdf (211.73 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-02403058 , version 1 (10-12-2019)

Identifiers

Cite

Pete Beckman, Emmanuel Jeannot, Swann Perarnau. Introduction to RADR 2019. IPDPSW 2019 - IEEE International Parallel and Distributed Processing Symposium Workshops, May 2019, Rio de Janeiro, Brazil. IEEE, pp.908-910, ⟨10.1109/IPDPSW.2019.00150⟩. ⟨hal-02403058⟩

Collections

CNRS INRIA INRIA2
19 View
54 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More