Analysis of Parallel Multidimensional Wave Digital Filtering Network on IBM Cell Broadband Engine

Chien Hsun Tseng

doi:10.1155/2014/793635

Abstract

As an alternative approach for the numerical integration of physical systems, the MDWDF technique has become of importance in the field of numerical analysis due to its attractive features, for example, massive parallelism and high accuracy both inherent in nature. In this study, speed-up efficiencies of a MDWDF network are studied for the linearized shallow water system, which plays an important role in fluid dynamics. To achieve the goal, the full parallelism of the MDWDF network is established in the first place based on the chained MD retiming technique. Following the implementation on the IBM Cell Broadband Engine (Cell/BE), excellent performance of the full parallel architecture is revealed. The IBM Cell/BE containing 1 power processor element (PPE) and 8 synergistic processor elements (SPEs) perfectly fits the architecture of the retimed MDWDF model. Empirical results have demonstrated that the full parallelized model with 8 processors (1PPE + 7SPEs) outperforms the other three models: partial right/left-loop retimed models and the full sequential model with 4× improvements for scheduled grids 51×51. In addition, for scheduled fine grids 201×201, the full parallel model is shown to possess significant performance over these models by up to 7× improvements.

Highlights

Physical system modeling is an important discipline in all fields as it enables the development of system simulation engines for physically realistic processes such as fluid flow, electrical, and acoustical phenomena
The most popular kind of models for multidimensional (MD) physical systems can be represented by sets of linear and/or nonlinear partial differential equations (PDEs) with properly imposed initial and boundary conditions
Adopting the full parallel architecture [11], in this study, the analysis of CPU runtime speed-up efficiencies is empirically implemented on the IBM Cell Broadband Engine (Cell/BE) to further improve performance of the MDWDF network representing the linearized shallow water (LSW) system, which plays an important role in fluid dynamics

Summary

Introduction

Physical system modeling is an important discipline in all fields as it enables the development of system simulation engines for physically realistic processes such as fluid flow, electrical, and acoustical phenomena. Making use of the WDF network paradigm and analogies with electrical networks, the MDWDF technique, draws a maximum advantage of essential physical properties of such systems, in particular of causality, passivity, stability, finite propagation velocity, and so forth, which can be all translated to the considered physical problems so as to preserve important relationships between variables [4, 5, 7–10] Proceeding in this way, it has the unique advantage of simultaneously offering the second-order accuracy, high robustness and fault tolerance, massive parallelism, full localness ( for taking into account arbitrary boundary conditions and shapes), and explicit or at least semiexplicit computability. For scheduled grids 201 × 201, the full parallel model is shown to possess significantly performance over these models by up to 7× improvements

Summary of 2D LSW System and Modeling Techniques

E4 EF4 Gc2 F4 G4 H4 C4

Hardware Experiment

E2 E3 E4

Conclusion

Conflict of Interests