Mapping Algorithms for Multiprocessor Tasks on Multi-Core Clusters

Dümmler, Jörg; Rauber, Thomas; Rünger, Gudula

doi:10.1109/icpp.2008.42

Cited by 11 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This looks like a task graph distribution problem, which is an important research field. For example, [3] proposes a scheduling algorithm on heterogeneous distributed architectures for static tasks modeled by Directed Acyclic Graph (DAG), and [4] introduces mapping strategies of hierarchically structured multiprocessor tasks on multicore clusters, which is our target architecture. However, our DACCOSIM FMU graph is not hierarchically structured, and is a DAG with very few task dependencies.…”

Section: Context and Related Workmentioning

confidence: 99%

Scaling of Distributed Multi-simulations on Multi-core Clusters

Dad¹,

Vialle²,

Caujolle³

et al. 2016

2016 IEEE 25th International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE)

View full text Add to dashboard Cite

DACCOSIM is a multi-simulation environment for continuous time systems, relying on FMI standard, making easy the design of a multi-simulation graph, and specially developed for multi-core PC clusters, in order to achieve speedup and size up. However, the distribution of the simulation graph remains complex and is still the responsibility of the simulation developer. This paper introduces DACCOSIM parallel and distributed architecture, and our strategies to achieve efficient multi-simulation graph distribution on multi-core clusters. Some performance experiments on two clusters, running up to 81 simulation components (FMU) and using up to 16 multi-core computing nodes, are shown. Performances measured on our faster cluster exhibit a good scalability, but some limitations of current DACCOSIM implementation are discussed.

show abstract

Section: Context and Related Workmentioning

confidence: 99%

Scaling of Distributed Multi-simulations on Multi-core Clusters

Dad¹,

Vialle²,

Caujolle³

et al. 2016

2016 IEEE 25th International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE)

View full text Add to dashboard Cite

show abstract

“…where: m indicates the number of processor nodes; SL(P i ) indicates the earliest possible completion time of all tasks which have been assigned to the processor P i . SL(P i )= max(SL(C ik ))，k=1,2,3,…,n (2) where: n indicates the number of processing cores in P i ; SL(C ik ) indicates the earliest possible completion time of all tasks which have been assigned to the processing core k in processor P i .…”

Section: Is the Number Of Successor Tasks Of T I ;mentioning

confidence: 99%

“…In the parallel programming model, the programming model based on message passing (such as MPI, PVM) is still a mainstream [2] . Parallel program based on the model contains multiple task(processes), and communication between tasks is by message passing.…”

Section: Introductionmentioning

confidence: 99%

A Task Scheduling Algorithm for Multi-Core-Cluster Systems

Geng¹,

Xu²,

Fu³

et al. 2012

JCP

View full text Add to dashboard Cite

The quantity of cores on one chip increases rapidly with the development of multi-core technology, which has led to more complex structure of cluster system and greatly increasing number of tasks. In order to schedule tasks in multi-core-cluster systems efficiently, a task schedule model based on the directed acyclic graph(DAG) is built, and then a algorithm based on task duplication is proposed. The algorithm is composed of two steps of operations, in which the processes are assigned to processor nodes in the first step and the threads in processes are assigned to core nodes in the second step respectively. The time complexity of this algorithm is less than similar algorithms. For the algorithm, minimization scheduling length is the primary objective, and keeping load balancing between processing nodes is secondary objectives. It can be seen through comparison with correlative work that the algorithm has advantages in scheduling length; furthermore, while the ratio of total communication cost and total computation cost in the task schedule model becomes larger, the advantage of this algorithm is more obvious.

show abstract

“…Most of those existing M-task scheduling algorithms are defined for homogeneous systems or distributed grid-like systems. Several strategies for mapping M-task applications on heterogeneous multi-core platforms have been presented in [5]. In this article, we extend this approach and propose a combined scheduling and mapping algorithm which is aware of the heterogeneity of multi-core systems.…”

Section: Introductionmentioning

confidence: 98%

Scalable computing with parallel tasks

Dümmler

Rauber

Rünger

2009

Proceedings of the 2nd Workshop on Many-Task Computing on Grids and Supercomputers

Self Cite

View full text Add to dashboard Cite

Recent and future parallel clusters and supercomputers use SMPs and multi-core processors as basic nodes, providing a huge amount of parallel resources. These systems often have hierarchically structured interconnection networks combining computing resources at different levels, starting with the interconnect within multi-core processors up to the interconnection network combining nodes of the cluster or supercomputer. The challenge for the programmer is that these computing resources should be utilized efficiently by exploiting the available degree of parallelism of the application programs and by structuring the application in a way which is sensitive to the heterogeneous interconnect.In this article, we present an approach to structure the computations of an application as parallel tasks which can interact with other parallel tasks in communication phases. In particular, we consider how these parallel tasks can be mapped onto the computing resources provided by parallel clusters or supercomputers. We show that the scalability can be significantly improved by a suitable task-based organization and a corresponding structuring of the communication within tasks as well as between tasks. We evaluate the impact of different mappings of tasks to cores for different application programs on a variety of parallel machines.

show abstract

Mapping Algorithms for Multiprocessor Tasks on Multi-Core Clusters

Cited by 11 publications

References 16 publications

Scaling of Distributed Multi-simulations on Multi-core Clusters

Scaling of Distributed Multi-simulations on Multi-core Clusters

A Task Scheduling Algorithm for Multi-Core-Cluster Systems

Scalable computing with parallel tasks

Contact Info

Product

Resources

About