An MPI-Stream Hybrid Programming Model for Computational Clusters

Mancini, Emilio; Marsh, Gregory J.; Panda, Dhabaleswar K.

doi:10.1109/ccgrid.2010.33

Cited by 11 publications

(3 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Mancini et al [33] propose a hybrid approach of Stream and MPI programming models, and use a whole MPI program as a building block in a stream application to improve processing speeds of computational units in hetrogeneous computational systems.…”

Section: Related Workmentioning

confidence: 99%

A Parallel Runtime Framework for Communication Intensive Stream Applications

Muralidharan

Casey

Gregg

2013

2013 12th IEEE International Conference on Trust, Security and Privacy in Computing and Communications

View full text Add to dashboard Cite

Stream applications are often limited in their performance by their underlying communication system. A typical implementation relies on the operating system to handle the majority of network operations. In such cases, the communication stack, which was not designed to handle tremendous amounts of data, acts as a bottleneck and restricts the performance of the application. In this paper, we propose a parallel runtime framework that integrates the communication operations with stream applications, and provides a common parallel processing engine that can execute both the communication and computation operations in parallel on multicore processors. We place an emphasis on the low-level details required to implement such a framework, but also provide some guidelines on how an application programmer can employ the framework. Our runtime system uses a set of operations represented as filters to perform the relevant computations on the data stream. Filters that handle the application specific operations are categorized as computation filters and those that transform data to and from network devices are classified as communication filters. Computation filters are designed by the user and are specific to the application. Communication filters are provided by the runtime system and are built using system software that allows direct access to network hardware. Such system software allows the network operations to be performed by the runtime system in parallel, leading to better communication performance. Applications that are designed for this framework are built by constructing application specific computation filters and then connecting them to the communication filters provided by the runtime system. This abstracts the low-level programming of network adapters and protocols by the application developer, making it easier to build stream applications that take advantage of the improved communication performance. Moreover, by dynamically replicating and statically scheduling such filters on the given multicore architecture, it is possible for the runtime system to process multiple data streams in parallel. We are able to parallelize stream applications and achieve speedups of more than a factor of eight in all the applications we tested. The results show that our system scales to as many parallel processes as there are cores on our computer, and achieves speedups of more than a factor of ten in some cases compared to sequential implementations.

show abstract

Section: Related Workmentioning

confidence: 99%

A Parallel Runtime Framework for Communication Intensive Stream Applications

Muralidharan

Casey

Gregg

2013

2013 12th IEEE International Conference on Trust, Security and Privacy in Computing and Communications

View full text Add to dashboard Cite

show abstract

“…Particularly, in most cases, an implementation of MPI is used and certain efforts have tried to integrate GPU programming environments and MPI in a unified development platform [14].…”

Section: Related Workmentioning

confidence: 99%

Accelerating data clustering on GPU-based clusters under shared memory abstraction

Karantasis

Polychronopoulos

Dimitrakopoulos

2010

2010 IEEE International Conference on Cluster Computing Workshops and Posters (CLUSTER WORKSHOPS)

View full text Add to dashboard Cite

Many-core graphics processors are playing today an important role in the advancements of modern highly concurrent processors. Their ability to accelerate computation is being explored under several scientific fields. In the current paper we present the acceleration of a widely used data clustering algorithm, K-means, in the context of high performance GPU clusters. As opposed to most related implementation efforts that use MPI to port their target applications on a GPU cluster, our implementation follows the Software Distributed Shared Memory (SDSM) paradigm in order to distribute information and computation across the accelerator cluster. In order to investigate the efficiency of a programming model that offers shared memory abstraction on GPU clusters we present two implementations, one that is based on a SDSM implementation of OpenMP and another that utilizes the Pleiad cluster middleware on top of the Java platform. The first results show that such an implementation is feasible in order to accelerate a broad category of large scale, data intensive applications, among which K-means is a characteristic case.

show abstract

“…We can generate safe MPI programs using not only fixed topologies such as pipelines or stencils, but also any well-formed Pabble protocols. As a portable standard, MPI is being adapted as a common interface to different kinds of programming models, including FPGAs [32], stream programming [23] and fault tolerant [13]. General MPI applications exhibit more complex communication patterns than well-known, connected topologies found in scientific computing.…”

Section: Introductionmentioning

confidence: 99%

Protocols by Default

Coutinho

Yoshida

2015

Lecture Notes in Computer Science

View full text Add to dashboard Cite

This paper presents a code generation framework for typesafe and deadlock-free Message Passing Interface (MPI) programs. The code generation process starts with the definition of the global topology using a protocol specification language based on parameterised multiparty session types (MPST). An MPI parallel program backbone is automatically generated from the global specification. The backbone code can then be merged with the sequential code describing the application behaviour, resulting in a complete MPI program. This merging process is fully automated through the use of an aspect-oriented compilation approach. In this way, programmers only need to supply the intended communication protocol and provide sequential code to automatically obtain parallelised programs that are guaranteed free from communication mismatch, type errors or deadlocks. The code generation framework also integrates an optimisation method that overlaps communication and computation, and can derive not only representative parallel programs with common parallel patterns (such as ring and stencil), but also distributed applications from any MPST protocols. We show that our tool generates efficient and scalable MPI applications, and improves productivity of programmers. For instance, our benchmarks involving representative parallel and application-specific patterns speed up sequential execution by up to 31 times and reduce programming effort by an average of 39%.

show abstract

An MPI-Stream Hybrid Programming Model for Computational Clusters

Cited by 11 publications

References 18 publications

A Parallel Runtime Framework for Communication Intensive Stream Applications

A Parallel Runtime Framework for Communication Intensive Stream Applications

Accelerating data clustering on GPU-based clusters under shared memory abstraction

Protocols by Default

Contact Info

Product

Resources

About