International Journal of Education and Management Engineering (IJEME)
IJEME Vol. 1, No. 3, 29 Sep. 2011
Cover page and Table of Contents: PDF (size: 157KB)
Performance Analysis for Heterogeneous & Reconfigurable Computing Based on Scheduling
Full Text (PDF, 157KB), PP.31-38
Views: 0 Downloads: 0
Author(s)
Yiming Tan
1
Guosun Zeng
1
Shuixia Hao
1
Index Terms
Component, heterogeneous & reconfigurable computing, performance analysis, heterogeneity matching matrix, reconfigurability coupling matrix
Abstract
Right now, heterogeneous & reconfigurable computing is a research hot in the area of high performance computing. Due to the heterogeneity of application tasks and reconfigurability of system architecture, performance analysis for heterogeneous & reconfigurable computing becomes rather difficult. Unfortunately, the existing techniques and methods are no longer suitable for use. This paper presents a performance analysis method based on task scheduling. It builds on system architecture model and task model of heterogeneous & reconfigurable computing. By making use of heterogeneity matching matrix and reconfigurability coupling matrix we achieve optimal selection and matching between computational tasks and processing units. Through task scheduling algorithm, the completion time of application task run on heterogeneous & reconfigurable computing system can be calculated. Finally, we carry out case study.
Cite This Paper
Yiming Tan,Guosun Zeng,Shuixia Hao,"Performance Analysis for Heterogeneous & Reconfigurable Computing Based on Scheduling", IJEME, vol.1, no.3, pp.31-38, 2011. DOI: 10.5815/ijeme.2011.03.05
Reference
[1]A.A. Khokhar, V.K. Prasanna, M.E. Shaaban. Heterogeneous computing: Challenges and opportunities. IEEE Computer, 1993, 26(6):18–27.
[2]K.J.Barker, K.Davis, A.Hoisie. Entering the petaflop era: the architecture and performance of Roadrunner. Proceedings of the ACM/IEEE conference on Supercomputing, 2008, pp15-21.
[3]K.Compton , S.Hauck. Reconfigurable Computing: A Survey of Systems and Software. ACM Computing Surveys, 2002, 34(2):171-210.
[4]F. Thoma, M.Kuhnle, P. Bonnot. MORPHEUS: Heterogeneous Reconfigurable Computing. Proceedings of 17th International Conference on Field Programmable Logic and Applications (FPL07), 2007, pp409–414.
[5]Zeng Guosun, Lu Xinda. Automatically Extracting The SIMD/MIMD Heterogeneity Hiding A Program [J].Journal of Computer Research & Development, 2000, 37(11):1397-1403(in Chinese)
[6]V.E.Bazterraa, M.Cumaa, M.B.Ferraroa. A general Framework to Understand Parallel Performance in Heterogeneous, Clusters: Analysis of A New adaptive Parallel Genetic Algorithm. Journal of Parallel and Distributed Computing, 2005, 65(1): 48-57.
[7]Yuan Lulai, Zeng Guosun. Dynamic Level Scheduling Based on Trust Model in Grid Computing [J]. Chinese Journal of Computers, 2006, 29(7): 1217-1224(in Chinese)
[8]A.Pombortsis, E.Papaefstathiou, A.Veglis. PASE: A performance analysis simulation environment. Simulation Practice and Theory, 1994, 2(1):43-59.
[9]J.Dongarra, A.Malony, S.MooreS. Performance Instrumentation and Measurement for Terascale Systems. Proceedings of the ICCS Conference (LNCS 2660), 2003, pp53–62.
[10]V.S.Adve, M.K.Vernon. Parallel program performance prediction using deterministic task graph analysis. ACM Transactions on Computer Systems, 2004, 22 (1): 94-136.
[11]M.C.Smith , G.D.Peterson. Parallel application performance on shared high performance reconfigurable computing resources. Performance Evaluation, 2005, 60(1-4):107-125.
[12]B.Javadi, M.K.Akbari, J.H.Abawajy. A performance model for analysis of heterogeneous multi-cluster systems. Journal of Parallel Computing, 2006, 32 (11-12):831-851.