Argobots: A Lightweight Low-Level Threading and Tasking Framework

Seo, Sangmin; Amer, Amer; Balaji, Pavan; Bordage, Cyril; Bosilca, George; Brooks, Alex; Carns, Philip; Castelló, Adrián; Genet, Damien; Hérault, Thomas; Iwasaki, Shinsuke; Jindal, Prateek; Kalé, Laxmikant V.; Krishnamoorthy, Sriram; Lifflander, Jonathan; Lu, Huiwei; Meneses, Esteban; Snir, Marc; Sun, Youwen; Taura, Kenjiro; Beckman, Pete

doi:10.1109/tpds.2017.2766062

Cited by 95 publications

(69 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In addition to languages and extensions, industry-standard and wellsupported parallel libraries based on task parallelism have emerged, such as Intel Cilk Plus [23] or Intel TBB [28]. There are also runtimes specifically designed to improve shared memory performance of existing language extensions, such as Qthreads [27] or Argobots [24]; this topic is of significant importance, considering the increase in many-core processors in recent years and, consequently, the importance of efficient lightweight runtimes. Task-based environments for heterogeneous hardware have also naturally developed with the emergence of accelerator and GPU computing; StarPU [2] is an example of such an environment.…”

Section: Stefano Markidis Markidis@kthsementioning

confidence: 99%

A taxonomy of task-based parallel programming technologies for high-performance computing

et al. 2018

View full text Add to dashboard Cite

Task-based programming models for shared memory-such as Cilk Plus and OpenMP 3-are well established and documented. However, with the increase in parallel, many-core, and heterogeneous systems, a number of research-driven projects have developed more diversified task-based support, employing various programming and runtime features. Unfortunately, despite the fact that dozens of different task-based systems exist today and are actively used for parallel and high-performance computing (HPC), no comprehensive overview or classification of task-based technologies for HPC exists. In this paper, we provide an initial task-focused taxonomy for HPC technologies, which covers both programming interfaces and runtime mechanisms. We

show abstract

Section: Stefano Markidis Markidis@kthsementioning

confidence: 99%

A taxonomy of task-based parallel programming technologies for high-performance computing

et al. 2018

View full text Add to dashboard Cite

show abstract

“…In the field of LWT libraries, the works in [6], [9], [12], [13], [14] introduce distinct LWT definitions, discuss implementation details, and analyze performance. The work in [24] conducts an analysis of different LWT solutions from the semantic point of view and evaluates their performance.…”

Section: Related Workmentioning

confidence: 99%

“…Compared with those, ConverseThreads [9] and Nanos++ [10] support a specific high-level PM; Charm++ [11] and OmpSs [5], respectively. There are also general-purpose solutions such as MassiveThreads [12], Qthreads [13], and Argobots [14]. The Generic Lightweight Threads (GLT) API [15] is an effort to unify these LWT solutions under a unique PM to foster productivity and portability with a negligible overhead that is demonstrated in [16].…”

Section: Introductionmentioning

confidence: 99%

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

Castelló

Seo

Mayo

et al. 2017

2017 46th International Conference on Parallel Processing (ICPP)

Self Cite

View full text Add to dashboard Cite

Abstract-OpenMP is the de facto standard application programming interface (API) for on-node parallelism. The most popular OpenMP runtimes rely on POSIX threads (pthreads) implementations that offer an excellent performance for coarsegrained parallelism and match perfectly with the current hardware. However, a recent trend in runtimes/applications points in the direction of leveraging massive on-node parallelism in conjunction with fine-grained and dynamic scheduling paradigms. It has been demonstrated that lightweight thread (LWT) solutions are more appropriate for these new parallel paradigms. We have developed GLTO, an OpenMP implementation over the recentlyemerged Generic Lightweight Threads (GLT) API. GLT exports a common API for LWT libraries that offers the possibility of running the same application over different native LWT solutions. In this paper we use GLTO to analyze different scenarios where OpenMP implementations may benefit from the use of either LWT or pthreads. Our study reveals that none of the threading approaches obtains the best performance in all the scenarios, but that there are important gaps among them.

show abstract

“…Different user-level thread (ULT) and tasklet libraries have been implemented in the past, such as Windows Fibers [14], Solaris Threads [2], ConverseThreads [13], Nanos++ [8], MassiveThreads [15], Qthreads [20], and Argobots [16] were the last three LWT solutions are compared. These solutions demonstrate semantic and performance benefits over the classic POSIX threads [3].…”

Section: Introductionmentioning

confidence: 99%

GLT: A Unified API for Lightweight Thread Libraries

Castelló

Seo

Mayo

et al. 2017

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. In recent years, several lightweight thread (LWT) libraries have emerged to tackle exascale challenges. These offer programming models (PMs) based on user-level threads and incorporate their own lightweight mechanisms. However, each library proposes its own PM, exposing different semantics and hindering portability. To address this drawback, we have designed Generic Lightweight Thread (GLT), an application programming interface that frames the functionality of the most popular LWT libraries for high-performance computing under a single PM. We implement GLT on top of Argobots, MassiveThreads, and Qthreads. We provide GLT as a dynamic library, as well as in the form of a static version based on macro preprocessing resolution to reduce overhead. This paper discusses the GLT PM and demonstrates its minimal performance impact.

show abstract

Argobots: A Lightweight Low-Level Threading and Tasking Framework

Cited by 95 publications

References 36 publications

A taxonomy of task-based parallel programming technologies for high-performance computing

A taxonomy of task-based parallel programming technologies for high-performance computing

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

GLT: A Unified API for Lightweight Thread Libraries

Contact Info

Product

Resources

About