A Review of Lightweight Thread Approaches for High Performance Computing

Abstract

Abstract-High-level, directive-based solutions are becoming the programming models (PMs) of the multi/many-core architectures. Several solutions relying on operating system (OS) threads perfectly work with a moderate number of cores. However, exascale systems will spawn hundreds of thousands of threads in order to exploit their massive parallel architectures and thus conventional OS threads are too heavy for that purpose. Several lightweight thread (LWT) libraries have recently appeared offering lighter mechanisms to tackle massive concurrency. In order to examine the suitability of LWTs in high-level runtimes, we develop a set of microbenchmarks consisting of commonlyfound patterns in current parallel codes. Moreover, we study the semantics offered by some LWT libraries in order to expose the similarities between different LWT application programming interfaces. This study reveals that a reduced set of LWT functions can be sufficient to cover the common parallel code patterns and that those LWT libraries perform better than OS threads-based solutions in cases where task and nested parallelism are becoming more popular with new architectures.

GLTO: On the Adequacy of Lightweight Thread Approaches for OpenMP Implementations

“…The work in [24] conducts an analysis of different LWT solutions from the semantic point of view and evaluates their performance.…”

GLT: A Unified API for Lightweight Thread Libraries

“…The library choices are based on the work presented in [9], where a set of LWT implementations was reviewed, from the semantic point of view, using a set of OpenMP microbenchmarks.…”

Unification of Lightweight Thread Solutions and their Application in High Performance Programming

“…The first step in designing a unified API was to review the existing threading solutions in order to extract common features from their PMs [33]. In this paper, threading libraries are decomposed and each PM is deeply analyzed.…”

“…Journal [33] Castelló, A., Mayo, R., Seo, S., Balaji, P., Quintana-Ortí, E. S., Peña, A. J. analysis of lightweight thread libraries for high-performance computing. Submitted to IEEE Transactions on Parallel and Distributed Systems (TPDS) (2018).…”

“…En esta sección se listan las publicaciones relacionadas con cada contribución y se clasifican como directamente relacionadas con el contenido de la tesis, indirectamente relacionadas y no relacionadas. El primer paso en el diseño de una interfaz común era analizar las soluciones existentes de hilos para poder extraer las características comunes de sus modelos de programación [33]. En este artículo las bibliotecas de hilos se analizan en profundidad.…”