Extending high performance Fortran for the support of unstructured computations

Müller, Andreas; Rühl, Roland

doi:10.1145/224538.224552

Cited by 24 publications

(6 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Existing parallel programming languages or models for distributed memory system provide various features to describe parallel programs and to execute them efficiently. For instance, XMP provides features similar to both CAF and HPF [19], allowing users to use either global or local view programming models, and providing easy to program functionality through the use of compiler directives for parallel functionality. Likewise, OpenMPD provided easy to program directives based parallelisation for message passing functionality, extending an OpenMP like approach to a distributed memory supercomputer.…”

Section: Related Workmentioning

confidence: 99%

MDMP: Managed Data Message Passing

Jackson¹,

Strand

2013

2013 15th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing

View full text Add to dashboard Cite

MDMP is a new parallel programming approach that aims to provide users with an easy way to add parallelism to programs, optimise the message passing costs of traditional scientific simulation algorithms, and enable existing MPIbased parallel programs to be optimised and extended without requiring the whole code to be re-written from scratch. MDMP utilises a directives based approach to enable users to specify what communications should take place in the code, and then implements those communications for the user in an optimal manner using both the information provided by the user and data collected from instrumenting the code and gathering information on the data to be communicated. In this paper we present the basic concepts and functionality of MDMP and discuss the performance that can be achieved using our prototype implementation of MDMP on some simple benchmark cases.

show abstract

Section: Related Workmentioning

confidence: 99%

MDMP: Managed Data Message Passing

Jackson¹,

Strand

2013

2013 15th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing

View full text Add to dashboard Cite

show abstract

“…However, since these works focus on array-based data parallel programs, they do not need to consider dynamic computation partitioning and data contention between threads. High Performance Fortran was extended for dynamic data distribution to parallelize array-based unstructured computations [Müller and Rühl 1995]. Several works [wei Liao et al 2006;Gordon et al 2006] have focused on data and task partitioning for stream programs.…”

Section: Related Workmentioning

confidence: 99%

PLDS

Feng

Lin

Gupta

2012

ACM Trans. Archit. Code Optim.

View full text Add to dashboard Cite

Recently, parallelization of computations in the presence of dynamic data structures has shown promising potential. In this paper, we present PLDS, a system for easily expressing and efficiently exploiting parallelism in computations that are based on dynamic linked data structures. PLDS improves the execution efficiency by providing support for data partitioning and then distributing computation across threads based on the partitioning. Such computations often require the use of speculation to exploit dynamic parallelism. PLDS supports a conditional speculation mechanism that reduces the cost of speculation. PLDS can be employed in the context of different forms of parallelism, which to cover a wide range of parallel applications. PLDS provides easy-to-use compiler directives, using enabling programmers to choose from among a variety of data partitionings to distribute computation across threads in a partitioning-sensitive fashion, and to use conditional speculation when required. We evaluate our implementation of PLDS using ten benchmarks, of which six are parallelized using speculation. PLDS achieves 1.3x-6.9x speedups on an 8-core machine.

show abstract

“…did not arise. Müller and Rühl proposed an extension to High Performance Fortran that allowed for its alignment and distribution idioms to be applied to irregular structures but their focus was on sparse arrays [20]. Recent work by Gordon et al has focused on exploiting data and task parallelism in stream programs [6].…”

Section: Related Work and Conclusionmentioning

confidence: 99%

Optimistic parallelism benefits from data partitioning

Kulkarni

Pingali

Ramanarayanan

et al. 2008

SIGOPS Oper. Syst. Rev.

View full text Add to dashboard Cite

Recent studies of irregular applications such as finite-element mesh generators and data-clustering codes have shown that these applications have a generalized data parallelism arising from the use of iterative algorithms that perform computations on elements of worklists. In some irregular applications, the computations on different elements are independent. In other applications, there may be complex patterns of dependences between these computations.The Galois system was designed to exploit this kind of irregular data parallelism on multicore processors. Its main features are (i) two kinds of set iterators for expressing worklist-based data parallelism, and (ii) a runtime system that performs optimistic parallelization of these iterators, detecting conflicts and rolling back computations as needed. Detection of conflicts and rolling back iterations requires information from class implementors.In this paper, we introduce mechanisms to improve the execution efficiency of Galois programs: data partitioning, data-centric work assignment, lock coarsening, and over-decomposition. These mechanisms can be used to exploit locality of reference, reduce mis-speculation, and lower synchronization overhead. We also argue that the design of the Galois system permits these mechanisms to be used with relatively little modification to the user code. Finally, we present experimental results that demonstrate the utility of these mechanisms.

show abstract

Extending high performance Fortran for the support of unstructured computations

Cited by 24 publications

References 20 publications

MDMP: Managed Data Message Passing

MDMP: Managed Data Message Passing

PLDS

Optimistic parallelism benefits from data partitioning

Contact Info

Product

Resources

About