Efficient Interprocedural Data Placement Optimisation in a Parallel Library

Beckmann, Olav

doi:10.1007/3-540-49530-4_9

Cited by 11 publications

(7 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of parallelization, Kennedy and Kremer [30], Anderson et al [6], and Eggers and Jeremiassen [26] used data transformation to improve locality for parallel programs. Beckman and Kelly studied run-time data layout selection in a parallel library [9]. The goal of most of these techniques is to improve data reuse within a single array.…”

Section: Related Work In Improving Cache Spatial Localitymentioning

confidence: 99%

Improving effective bandwidth through compiler enhancement of global cache reuse

Ding

Kennedy²

2004

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

Section: Related Work In Improving Cache Spatial Localitymentioning

confidence: 99%

Improving effective bandwidth through compiler enhancement of global cache reuse

Ding

Kennedy²

2004

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

“…Results reported in [6] for the conjugate gradient algorithm on an ethernet-connected Linux cluster show a reduction in communication time of 15%-50%, although the impact on overall execution time is smaller.…”

Section: The Communication Fusion Library As a Component Modelmentioning

confidence: 98%

Generative and Adaptive Methods in Performance Programming

Beckmann

2005

Parallel Process. Lett.

Self Cite

View full text Add to dashboard Cite

Performance programming is characterized by the need to structure software components to exploit the context of use. Relevant context includes the target processor architecture, the available resources (number of processors, network capacity), prevailing resource contention, the values and shapes of input and intermediate data structures, the schedule and distribution of input data delivery, and the way the results are to be used. This paper concerns adapting to dynamic context: adaptive algorithms, malleable and migrating tasks, and application structures based on dynamic component composition. Adaptive computations use metadata associated with software componentsperformance models, dependence information, data size and shape. Computation itself is interwoven with planning and optimizing the computation process, using this metadata. This reflective nature motivates metaprogramming techniques. We present a research agenda aimed at developing a modelling framework which allows us to characterize both computation and dynamic adaptation in a way that allows systematic optimization.

show abstract

“…in U, due to the memory reference U In our prior work [6][7][8][9][10], component metadata describes data placement constraints. In this framework, component dependence metadata captures the available flexibility in execution order.…”

Section: Example: Multiblock Jacobimentioning

confidence: 99%

“…The background to this is our body of work in cross-component data placement optimisation for regular, data-parallel programs [6][7][8][9]. We have developed a good understanding of the parameters that determine both the complexity and the accuracy of optimisation algorithms in this problem domain [10].…”

Section: Introductionmentioning

confidence: 99%

Themis: Component Dependence Metadata in Adaptive Parallel Applications

Beckmann

Field

Baden

2001

Parallel Process. Lett.

Self Cite

View full text Add to dashboard Cite

This paper describes THEMIS, a programming model and run-time library being designed to support cross-component performance optimization through explicit manipulation of the computation's iteration space at run-time. Each component is augmented with "component dependence metadata", which characterizes the constraints on its execution order, data distribution and memory access order. We show how this supports dynamic adaptation of each component to exploit the available resources, the context in which its operands are generated, and results are used, and the evolution of the problem instance. Using a computational fluid dynamics visualization example as motivation, we show how component dependence metadata provides a framework in which a number of interesting optimizations become possible. Examples include data placement optimization, loop fusion, tiling, memoization, checkpointing and incrementalization.

show abstract

Efficient Interprocedural Data Placement Optimisation in a Parallel Library

Cited by 11 publications

References 9 publications

Improving effective bandwidth through compiler enhancement of global cache reuse

Improving effective bandwidth through compiler enhancement of global cache reuse

Generative and Adaptive Methods in Performance Programming

Themis: Component Dependence Metadata in Adaptive Parallel Applications

Contact Info

Product

Resources

About