Single Data Copying for MPI Communication Optimization on Shared Memory System

Miao, Qiankun; Sun, Guoqiang; Shan, Jiulong; Chen, Guoliang

doi:10.1007/978-3-540-72584-8_93

Cited by 2 publications

(2 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For Xeon Phi, Potluri et al optimised MVAPICH2 by leveraging the Symmetric Communications InterFace, a low‐level API that allows to control the DMA engines. Miao et al proposed a single‐copy method to take advantage of shared memory architectures, avoiding the system buffer. The proposal of Trahay et al relies on a multithreaded communication engine to offload small messages .…”

Section: Related Workmentioning

confidence: 99%

MPI and UPC broadcast, scatter and gather algorithms in Xeon Phi

Mallón

Taboada

Koesterke³

2015

Concurrency and Computation

View full text Add to dashboard Cite

SUMMARYAccelerators have revolutionised the high performance computing (HPC) community. Despite their advantages, their very specific programming models and limited communication capabilities have kept them in a supporting role of the main processors. With the introduction of Xeon Phi, this is no longer true, as it can be programmed as the main processor and has direct access to the InfiniBand network adapter. Collective operations play a key role in many HPC applications. Therefore, studying its behaviour in the context of manycore coprocessors has great importance. This work analyses the performance of different algorithms for broadcast, scatter and gather, in a large-scale Xeon Phi supercomputer. The algorithms evaluated are those available in the reference message passing interface (MPI) implementation for Xeon Phi (Intel MPI), the default algorithm in an optimised MPI implementation (MVAPICH2-MIC), and a new set of algorithms, developed by the authors of this work, designed with modern processors and new communication features in mind. The latter are implemented in Unified Parallel C (UPC), a partitioned global address space language, leveraging one-sided communications, hierarchical trees and message pipelining. This study scales the experiments to 15 360 cores in the Stampede supercomputer and compares the results to Xeon and hybrid Xeon + Xeon Phi experiments, with up to 19 456 cores.

show abstract

Section: Related Workmentioning

confidence: 99%

MPI and UPC broadcast, scatter and gather algorithms in Xeon Phi

Mallón

Taboada

Koesterke³

2015

Concurrency and Computation

View full text Add to dashboard Cite

show abstract

“…Following the first approach, Miao et al [20] proposed a single copying method to take advantage of shared memory architectures, avoiding the system buffer. The proposal of Trahay et al relies on a multithreaded communication engine to offload small messages [30].…”

Section: Related Workmentioning

confidence: 99%

Scalable PGAS collective operations in NUMA clusters

et al. 2014

View full text Add to dashboard Cite

The increasing number of cores per processor is turning manycore-based systems in pervasive. This involves dealing with multiple levels of memory in NUMA (Non Uniform Memory Access) systems and processor cores hierarchies, accessible via complex interconnects in order to dispatch the increasing amount of data required by the processing elements. The key for efficient and scalable provision of data is the use of collective communication operations that minimize the impact of bottlenecks. Leveraging one sided communications becomes more important in these systems, to avoid unnecessary synchronization between pairs of processes in collective operations implemented in terms of two sided point to point functions. This work proposes a series of algorithms that provide a good performance and scalability in collective operations, based on the use of hierarchical trees, overlapping one-sided communications, message pipelining and the available NUMA binding features. An implementation has been developed for Unified Parallel C (UPC), a PGAS (Partitioned Global Address Space) language, which presents a shared memory view across the nodes for programmability, while keeping private memory regions for performance. The performance evaluation of the proposed implementation, conducted on five representative systems (JuRoPA, JUDGE, Finis Terrae, SVG and Superdome), has shown generally good performance and scalability, even outperforming MPI in some cases, which confirms the suitability of the developed algorithms for manycore architectures.

show abstract

Single Data Copying for MPI Communication Optimization on Shared Memory System

Cited by 2 publications

References 7 publications

MPI and UPC broadcast, scatter and gather algorithms in Xeon Phi

MPI and UPC broadcast, scatter and gather algorithms in Xeon Phi

Scalable PGAS collective operations in NUMA clusters

Contact Info

Product

Resources

About