Expressing and Exploiting Multi-Dimensional Locality in DASH

Fuchs, Tobias; Fürlinger, Karl

doi:10.1007/978-3-319-40528-5_15

Cited by 6 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MPI ranks). As a basis we use the pattern concept as described in previous work [8]. The mapping of data onto global memory location is a three-step process: first, each element in a container is assigned to a location in the global index space.…”

Section: Data Placementmentioning

confidence: 99%

Portable Node-Level Parallelism for the PGAS Model

Jungblut

Fürlinger

2021

Int J Parallel Prog

Self Cite

View full text Add to dashboard Cite

The Partitioned Global Address Space (PGAS) programming model brings intuitive shared memory semantics to distributed memory systems. Even with an abstract and unifying virtual global address space it is, however, challenging to use the full potential of different systems. Without explicit support by the implementation node-local operations have to be optimized manually for each architecture. A goal of this work is to offer a user-friendly programming model that provides portable performance across systems. In this paper we present an approach to integrate node-level programming abstractions with the PGAS programming model. We describe the hierarchical data distribution with local patterns and our implementation, MEPHISTO, in C++ using two existing projects. The evaluation of MEPHISTO shows that our approach achieves portable performance while requiring only minimal changes to port it from a CPU-based system to a GPU-based one using a CUDA or HIP back-end.

show abstract

Section: Data Placementmentioning

confidence: 99%

Portable Node-Level Parallelism for the PGAS Model

Jungblut

Fürlinger

2021

Int J Parallel Prog

Self Cite

View full text Add to dashboard Cite

show abstract

“…Thus a fully 1. The figure is inspired by Fuchs and Fuerlinger [21]. Index Space 0" 1" 2" 3" 4" 5" 6" 7" 8" 9" 10" 11" 12" 13" 14" 15" 16" 17" 18" 19" 20" 21" 22" 23" Data Decomposition 0" 1" 2" 3" 4" 5" 6" 7" 8" 9" 10" 11" 12" 13" 14" 15" 16" 17" 18" 19" 20" 21" 22" 23"…”

Section: Terminologymentioning

confidence: 99%

Trends in Data Locality Abstractions for HPC Systems

Unat

Dubey

Hoefler

et al. 2017

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

Abstract-The cost of data movement has always been an important concern in high performance computing (HPC) systems. It has now become the dominant factor in terms of both energy consumption and performance. Support for expression of data locality has been explored in the past, but those efforts have had only modest success in being adopted in HPC applications for various reasons. However, with the increasing complexity of the memory hierarchy and higher parallelism in emerging HPC systems, locality management has acquired a new urgency. Developers can no longer limit themselves to low-level solutions and ignore the potential for productivity and performance portability obtained by using locality abstractions. Fortunately, the trend emerging in recent literature on the topic alleviates many of the concerns that got in the way of their adoption by application developers. Data locality abstractions are available in the forms of libraries, data structures, languages and runtime systems; a common theme is increasing productivity without sacrificing performance. This paper examines these trends and identifies commonalities that can combine various locality concepts to develop a comprehensive approach to expressing and managing data locality on future large-scale high-performance computing systems.

show abstract

“…This contradicts the driving idea behind PGAS: ease of programmability. The good news is, DASH is tackling these issues by providing automatic optimization for faster local data accesses (Zhou et al 2015), advanced data distribution schemes (Fuchs et al 2015), algorithm specific routines for pre-fetching and packing of data and automatic hierarchical units mapping. The following section provides a detailed illustration of these advance features.…”

Section: Experiences With Upcmentioning

confidence: 99%

“…We therefore provide high-level functions to automatically optimize data distribution for a given algorithm and vice-versa. These mechanism are described in detail in (Fuchs et al 2015). In this we present a classification of data distribution schemes based on well-defined properties in the three mapping stages of domain decomposition: partitioning, mapping, and memory layout.…”

Section: High-level Data Distribution Schemesmentioning

confidence: 99%

Effective use of the PGAS Paradigm: Driving Transformations and Self-Adaptive Behavior in DASH-Applications

Idrees,

Fuchs,

Glass

2016

Preprint

Self Cite

View full text Add to dashboard Cite

DASH is a library of distributed data structures and algorithms designed for running the applications on modern HPC architectures, composed of hierarchical network interconnections and stratified memory. DASH implements a PGAS (partitioned global address space) model in the form of C++ templates, built on top of DART -a run-time system with an abstracted tier above existing one-sided communication libraries.In order to facilitate the application development process for exploiting the hierarchical organization of HPC machines, DART allows to reorder the placement of the computational units. In this paper we present an automatic, hierarchical units mapping technique (using a similar approach to the Hilbert curve transformation) to reorder the placement of DART units on the Cray XC40 machine Hazel Hen at HLRS. To evaluate the performance of new units mapping which takes into the account the topology of allocated compute nodes, we perform latency benchmark for a 3D stencil code. The technique of units mapping is generic and can be be adopted in other DART communication substrates and on other hardware platforms.Furthermore, high-level features of DASH are presented, enabling more complex automatic transformations and optimizations in the future.

show abstract

Expressing and Exploiting Multi-Dimensional Locality in DASH

Cited by 6 publications

References 12 publications

Portable Node-Level Parallelism for the PGAS Model

Portable Node-Level Parallelism for the PGAS Model

Trends in Data Locality Abstractions for HPC Systems

Effective use of the PGAS Paradigm: Driving Transformations and Self-Adaptive Behavior in DASH-Applications

Contact Info

Product

Resources

About