论文信息 - Performance Analysis and Improvement of OpenMP on Software Distributed Shared Memory Systems

Performance Analysis and Improvement of OpenMP on Software Distributed Shared Memory Systems

In this paper, the performance of the portable OpenMP com- piler on SDSM JIAJIA is analyzed using SPEC OMPM2001 benchmark. The overheads of parallel execution have been investigated from the as- pects of thread management and task schedule, memory access and syn- chronization. To improve the performance, the page placement and data privatization techniques have been implemented for the optimization of the compiler.

Yang Xuejun | Huang Chun

[1] Rudolf Eigenmann,et al. Large System Performance of SPEC OMP2001 Benchmarks , 2002, ISHPC.

[2] Alan L. Cox,et al. TreadMarks: shared memory computing on networks of workstations , 1996 .

[3] Rudolf Eigenmann,et al. SPEComp: A New Benchmark Suite for Measuring Parallel Computer Performance , 2001, WOMPAT.

[4] Liviu Iftode,et al. Scope Consistency: A Bridge between Release Consistency and Entry Consistency , 1996, SPAA '96.

[5] David A. Padua,et al. Is OpenMP for grids ? , 2002, Proceedings 16th International Parallel and Distributed Processing Symposium.

[6] W. Hu,et al. JIA-JIA : An SVM System Based on A New Cache Coherence Protocol , 1999 .

[7] Nicholas P. Carter,et al. Processor mechanisms for software shared memory , 2000, ISHPC.

[8] Rudolf Eigenmann,et al. Towards OpenMP Execution on Software Distributed Shared Memory Systems , 2002, ISHPC.

[9] Willy Zwaenepoel,et al. OpenMP on Networks of Workstations , 1998, Proceedings of the IEEE/ACM SC98 Conference.