Titanium

Robert, Yves; Shende, S M; Malony, Allen D.; Morris, A.R.; Spear, W. S.; Biersdorff, Scott; Smith, Burton; Wang, Dali; Ricciuto, D. M.; Post, W.; Berry, Michael W.; Irigoin, François; Yelick, Katherine; Graham, Susan L.; Hilfinger, Paul; Bonachea, Dan; Su, Jimmy; Kamil, Amir; Datta, Kaushik; Colella, Phillip; Wen, Tong; Dongarra, Jack; Łuszczek, Piotr; Bhatelé, Abhinav; Freudenberger, Stefan M.; Diekert, Volker; Muscholl, Anca; Herlihy, Maurice; Moss, J. Eliot B.

doi:10.1007/978-0-387-09766-4_516

Cited by 2 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Namely, the private space is always the one that can be more efficiently accessed and the shared local space is accessible in very similar times, while the accesses to remote shared data are the most expensive ones by far. This model has been implemented in a wide variety of languages [2], [3], [4], [5], [6] and libraries [7].…”

Section: Introductionmentioning

confidence: 99%

“…Despite the advantages of PGAS approaches and the large set of semantics and mechanisms they provide, they still lack important features available in other parallel frameworks. This way, a common restriction we have found is that although PGAS environments exhibit a large variety of synchronization mechanisms, from full/empty bits [5] or synchronized blocks [6] to clocks [4] and locks [3], it is difficult to efficiently develop many algorithms whose tasks present complex patterns of dependencies using these approaches. A very effective mechanism to express this kind of algorithms is to rely on the implicit dependencies between the tasks that are given by the definition of their inputs and outputs.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Easy Dataflow Programming in Clusters with UPC++ DepSpawn

Fraguela

Andrade

2019

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

The Partitioned Global Address Space (PGAS) programming model is one of the most relevant proposals to improve the ability of developers to exploit distributed memory systems. However, despite its important advantages with respect to the traditional message-passing paradigm, PGAS has not been yet widely adopted. We think that PGAS libraries are more promising than languages because they avoid the requirement to (re)write the applications using them, with the implied uncertainties related to portability and interoperability with the vast amount of APIs and libraries that exist for widespread languages. Nevertheless, the need to embed these libraries within a host language can limit their expressiveness and very useful features can be missing. This paper contributes to the advance of PGAS by enabling the simple development of arbitrarily complex task-parallel codes following a dataflow approach on top of the PGAS UPC++ library, implemented in C++. In addition, our proposal, called UPC++ DepSpawn, relies on an optimized multithreaded runtime that provides very competitive performance, as our experimental evaluation shows.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Easy Dataflow Programming in Clusters with UPC++ DepSpawn

Fraguela

Andrade

2019

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

show abstract

High-performance dataflow computing in hybrid memory systems with UPC++ DepSpawn

Fraguela

Andrade

2021

J Supercomput

View full text Add to dashboard Cite

Dataflow computing is a very attractive paradigm for high performance computing, given its ability to trigger computations as soon as their inputs are available. UPC++ DepSpawn is a novel task-based library that supports this model in hybrid shared/distributed memory systems on top of a Partitioned Global Address Space (PGAS) environment. While the initial version of the library provided good results, it suffered from a key restriction that heavily limited its performance and scalability. Namely, each process had to consider all the tasks in the application rather than only those of interest to it, an overhead that naturally grows both with the number of processes and tasks in the system. In this paper this restriction is lifted, enabling our library to provide higher levels of performance. This way, in experiments using 768 cores the performance improved up to 40.1%, the average improvement being 16.1%.

show abstract

Titanium

Cited by 2 publications

References 0 publications

Easy Dataflow Programming in Clusters with UPC++ DepSpawn

Easy Dataflow Programming in Clusters with UPC++ DepSpawn

High-performance dataflow computing in hybrid memory systems with UPC++ DepSpawn

Contact Info

Product

Resources

About