Optimal Use of Mixed Task and Data Parallelism for Pipelined Computations

Subhlok, Jaspal; Vondran, Gary

doi:10.1006/jpdc.1999.1596

Cited by 49 publications

(25 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, one could envision extending our application model to address the situation in which each divisible load application consists of a set of tasks linked by dependencies. This would be an attractive extension of the mixed task and data parallelism approach (Chakrabarti, Demmel, and Yelick 1997;Subhlok and Vondran 2000;Braun et al 2001) The construction of these sets shows that there is an edge e k between V i and V j in G.…”

Section: Resultsmentioning

confidence: 99%

Steady-State Scheduling of Multiple Divisible Load Applications on Wide-Area Distributed Computing Platforms

Marchal

Yang

Casanova³

et al. 2006

The International Journal of High Performance Computing Applica

View full text Add to dashboard Cite

Divisible load applications consist of an amount of data and associated computation that can be divided arbitrarily into any number of independent pieces. This model is a good approximation of many real-world scientific applications, lends itself to a natural master-worker implementation, and has thus received a lot of attention. The critical issue of divisible load scheduling has been studied extensively in previous work. However, only a few authors have explored the simultaneous scheduling of multiple such applications on a distributed computing platform. We focus on this increasingly relevant scenario and make the following contributions. We use a novel and more realistic platform model that captures some of the fundamental network properties of grid platforms. We formulate the steady-state multi-application scheduling problem as a linear program that expresses a notion of fairness between applications. This scheduling problem is NP-complete and we propose several heuristics that we evaluate and compare via extensive simulation experiments. Our main finding is that some of our heuristics can achieve performance close to the optimal and we quantify the trade-offs between achieved performance and heuristic complexity.

show abstract

Section: Resultsmentioning

confidence: 99%

Steady-State Scheduling of Multiple Divisible Load Applications on Wide-Area Distributed Computing Platforms

Marchal

Yang

Casanova³

et al. 2006

The International Journal of High Performance Computing Applica

View full text Add to dashboard Cite

show abstract

“…Subholk and Vodran have presented a dynamic programming approach to determine the optimal division of dataand pipeline-parallelism in an application to meet throughput and latency constraints [18]. This work uses a slightly more restrictive programming model of a single pipeline of tasks, but demonstrates a more complete analytical analysis than that presented here or in StreamIt [13].…”

Section: Related Workmentioning

confidence: 99%

Block-Parallel Programming for Real-Time Embedded Applications

Black-Schaffer

Dally

2010

2010 39th International Conference on Parallel Processing

View full text Add to dashboard Cite

Abstract-Embedded media applications have traditionally used custom ASICs to meet their real-time performance requirements. However, the combination of increasing chip design cost and availability of commodity many-core processors is making programmable devices increasingly attractive alternatives. Yet for these processors to be successful in this role, programming systems are needed that can automate the task of mapping the applications to the tens-tohundreds of cores on current and future many-core processors, while simultaneously guaranteeing the real-time throughput constraints. This paper presents a block-parallel program description for embedded real-time media applications and automatic transformations including buffering and parallelization to ensure the program meets the throughput requirements. These transformations are enabled by starting with a high-level, yet intuitive, application description. The description builds on traditional stream programming structures by adding simple control and serialization constructs to enable a greater variety of applications. The result is an application description that provides a balance of flexibility and power to the programmer, while exposing the application structure to the compiler at a high enough level to enable useful transformations without heroic analysis.

show abstract

“…In [13], the authors propose a dynamic programming solution for the problem of minimizing the latency with a throughput constraint and present a near optimal solution to the problem of maximizing the throughout with a latency constraint on homogeneous platforms.…”

Section: Pipelined Data Parallel Tasksmentioning

confidence: 99%

Ocean-Atmosphere Modelization over the Grid

Caniou

Caron

Charrier

et al. 2008

2008 37th International Conference on Parallel Processing

View full text Add to dashboard Cite

Abstract:In this report, we tackle the problem of scheduling an Ocean-Atmosphere application used for climate prediction on the grid. An experiment is composed of several 1D-meshes of identical DAGs composed of parallel tasks. To obtain a good completion time, we divide groups of processors into sets each working on parallel tasks. The group sizes are chosen by computing the best makespan for several grouping possibilities. We improved this heuristic method by different means. The improvement yielding to the best makespan is the representation of the problem as an instance of the Knapsack problem. As this heuristic is firstly designed for homogeneous platforms, we present its adaptation to heterogeneous platforms. Simulations show improvements of the makespan up to 12%. Key-words: Grid computing, Ocean-Atmosphere application, SchedulingThis text is also available as a research report of the Laboratoire de l'Informatique du Parallélisme http://www.ens-lyon.fr/LIP.Modélisation Océan-Atmosphère sur la grille Résumé : Les besoins en calcul pour la simulation de modèle climatologique sont très importants. Afin d'améliorer les temps de calcul, nous nous intéressons ici au problème de l'ordonnancement de simulations du climat sur une grille de calcul. Plusieurs simulations couplées Océan-Atmosphère sont entreprises simultanément afin d'estimer statistiquement l'incertitude sur le réchauffement du climat du XXIe siècle. Notre expérience est donc composée de plusieurs chaînes de DAG (Directed Acyclic Graph) identiques composés de tâches parallèles. Afin de minimiser le temps d'exécution, nous divisons des groupes de processeurs en ensembles, chacun travaillant sur une tâche parallèle. Les tailles des groupes sont choisis en estimant le meilleur temps d'exécution en fonction de plusieurs scénarios de regroupement. Nous avons mis en oeuvre et amélioré une heuristique pour plate-forme homogène pour résoudre ce problème. L'heuristique offrant les meilleures performances est la solution s'inspirant du problème du sac à dos. Nous proposons ensuite une adaptation de cette heuristique dans le cadre de plates-formes hétérogènes. Enfin nous proposons quelques simulations qui montrent entre autres des gains allant jusqu'à 12%.

show abstract

Optimal Use of Mixed Task and Data Parallelism for Pipelined Computations

Cited by 49 publications

References 21 publications

Steady-State Scheduling of Multiple Divisible Load Applications on Wide-Area Distributed Computing Platforms

Steady-State Scheduling of Multiple Divisible Load Applications on Wide-Area Distributed Computing Platforms

Block-Parallel Programming for Real-Time Embedded Applications

Ocean-Atmosphere Modelization over the Grid

Contact Info

Product

Resources

About