Formulating the real cost of DSM-inherent dependent parameters in HPC clusters

The choice of an appropriate interprocess communication (IPC) mechanism is critical to the performance of distributed systems and parallel programs. There is however a trade-off between the accrued system performance and the imposed cost of the deployed IPC mechanism. Message passing (MP) and distributed shared memory (DSM) mechanisms have been extensively studied and compared, but the state of the art work on formulating the cost of DSM is not comprehensive. In this paper we distinguish between DSM-inherent and application-specific parameters that contribute most to the real cost of DSM, and focus on DSM-inherent parameters in high performance computing (HPC) clusters. The weights of each parameter's effectiveness on the cost, as well as the weights of their influence on each other are determined. The derived formula for the real cost of DSM is then calibrated by a clustering coefficient that varies with different sizes of clusters. We have used the principles of management and accounting sciences in calculating the real cost of DSM.

[1]  Rajkumar Buyya,et al.  Cluster Computing: High-Performance, High-Availability, and High-Throughput Processing on a Network of Computers , 2006, Handbook of Nature-Inspired and Innovative Computing.

[2]  H. Smalley.  The systems approach. , 1972, Hospitals.

[3]  Mohsen Sharifi,et al.  Evaluating the Effect of Inter Process Communication Efficiency on High Performance Distributed Scientific Computing , 2008, 2008 IEEE/IFIP International Conference on Embedded and Ubiquitous Computing.

[4]  Kourosh Gharachorloo,et al.  Towards transparent and efficient software distributed shared memory , 1997, SOSP.

[5]  M. Anthony,et al.  Advanced linear algebra , 2006 .

[6]  Nitin H. Vaidya,et al.  A cost model for distributed shared memory using competitive update , 1997, Proceedings Fourth International Conference on High-Performance Computing.

[7]  Rob F. Van der Wijngaart,et al.  Predicting cost/performance trade-offs for Whitney: a commodity computing cluster , 1998, Proceedings of the Thirty-First Hawaii International Conference on System Sciences.

[8]  Mohsen Sharifi,et al.  The Influence of Efficient Message Passing Mechanisms on High Performance Distributed Scientific Computing , 2008, 2008 IEEE International Symposium on Parallel and Distributed Processing with Applications.

[9]  William J. Bruns,et al.  Accounting and Management: Field Study Perspectives , 1987 .

[10]  Wilson C. Hsieh,et al.  The persistent relevance of IPC performance: new techniques for reducing the IPC penalty , 1993, Proceedings of IEEE 4th Workshop on Workstation Operating Systems. WWOS-III.

[11]  Michael Stumm,et al.  Algorithms implementing distributed shared memory , 1990, Computer.

[12]  James R. Larus,et al.  Sirocco: cost-effective fine-grain distributed shared memory , 1998, Proceedings. 1998 International Conference on Parallel Architectures and Compilation Techniques (Cat. No.98EX192).

[13]  Brett D. Fleisch,et al.  A Dynamic Coherence Protocol for Distributed Shared Memory Enforcing High Data Availability at Low Costs , 1996, IEEE Trans. Parallel Distributed Syst..

[14]  Astrid Kiehn.  An Operational Semantics for Shared Messaging Communication , 2007, Electron. Notes Theor. Comput. Sci..

[15]  Graham Morgan,et al.  Design and Evaluation of a WideArea Distributed Shared Memory Middleware , 2007 .

[16]  Jian S. Dai,et al.  Product Cost Estimation: Technique Classification and Methodology Review , 2006 .

[17]  S. K. Nandy,et al.  A complexity effective communication model for behavioral modeling of signal processing applications , 2003, DAC '03.

[18]  Dana S. Henry,et al.  A tightly-coupled processor-network interface , 1992, ASPLOS V.

[19]  Manu Konchady.  Parallel Computing Using Linux , 1998 .

[20]  Christian Kurmann,et al.  Cost/performance tradeoffs in network interconnects for clusters of commodity PCs , 2003 .

[21]  Galen C. Hunt,et al.  Shared memory computing on clusters with symmetric multiprocessors and system area networks , 2005, TOCS.

[22]  Nitin H. Vaidya,et al.  A cost-comparison approach for adaptive distributed shared memory , 1996, ICS '96.