Journal Press India^®

Author Details ( * ) denotes Corresponding author

A number of design techniques for multiprocessing architectures have been investigated as a result of various developments in IC processing and integrating technologies. While designing massively parallel computer systems, the choice of the interconnection network's topology is one of the many crucial design concerns. And as a result, there have been numerous proposals for interconnection networks in the literature, and a ton of study has been done on the creation and evaluation of these interconnection networks. However, the issue of integrating the processing components in multiprocessing parallel architectures in order to achieve great computational efficiency has not yet been fully solved. In order to effectively manage parallelism on an interconnection network, it is necessary to maximize a number of competing performance indicators, such as reducing communication and scheduling overheads and distributing workloads evenly. In order to reduce communication cost, load balancing entails distributing work to each processor in proportion to its performance. The assignment may be completed statically at compile time or dynamically at run time. Many load balancing policies boost system performance by using more processing power, memory, or a combination of the two. The present work is centered on implementation of two existing dynamic load balancing schemes –Sender Initiated Diffusion (SID) and Receiver Initiated Diffusion (RID) to the Linearly Extensible Multiprocessor (LEM) architecture The results achieved in the simulation are presented to evaluate the performance of LEM architecture.

Keywords

LEC; LET; SID; Load Balancing; Time; Load Imbalance Factor; Ideal Load; Linearly Extensible Triangle

1. Abdus Samad & Qasim Rafiq ―A Novel architecture LEC for Network Sever‖, Ph.D, 2010.


2. Rafiq M.Q., M. Ba-Ru-K & A.Samad, ― A Linearly Extensible Cube Network‖ sent for publication in IEEE, Jan. 2001


3. Rafiq M.Q., M. Ba-Ru-K & A.Samad, ― A Linearly Extensible Cube Network‖ sent for publication in IEEE, Jan. 2001


4. Abdallah   Boukerram,   Samira   Ait   Kaci   Azzou, “Implementation of Load Balancing Algorithm in a Grid Computing”, American Journal of Applied Sciences, 2006.


5. Manaullah, “Performance Evaluation of Multiprocessor Architectures”, Ph.D. thesis, Jamia Millia Islamia, 2002.


6. Abdus Samad, “Performance Evaluation of Linearly Extensible Multiprocessor Architectures for Networking”, Ph.D. thesis, Aligarh Muslim University, 2009.


7. M. Q. Rafiq, “Studies on the Performance Evaluation of a Linearly Extensible Multiprocessor Network”, Ph.D thesis, Univ. of Roorkee, 1995.


8. D.  Acker,  S.  Kulkarni,  “A  Dynamic  Load  Dispersion Algorithm for Load-Balancing in a Heterogeneous Grid System”, Sarnoff Symposium IEEE, pp 1- 5, 2007.


9. A.   Chhabra,   G.   Singh,   “Qualitative   Parametric Comparison of Load Balancing Algorithms in Distributed Computing Environment”, 14th International Conference on Advanced Computing and Communication, IEEE, pp 58– 61, 2006.


10. Neeraj K., “Simulation Study for Performance and Prediction of Parallel Computers,” BVICAM’s International Journal of  Information Technology, (BIJIT) , vol. 4, 2012


11. Abdus Samad, Jamshed Siddiqui “Properties and Performance of Cube-based Multiprocessor Architectures” IAJIT international journal, vol-3, 2014  


12. Marc H. Willebeek-LeMair and Anthony P. Reeves “Strategies for Dynamic Load Balancing on Highly Parallel Computers” IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, VOL. 4, NO. 9, SEPTEMBER 1993






 
13.  
14.  
15.  
16.