X. Zhang and Y. Yan, "Comparative modeling and evaluation of CC-NUMA and COMA on hierarchical ring architectures", IEEE Transactions on Parallel and Distributed Systems, Vol. 6, No. 12, 1995, pp. 1316-1331.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... to flatten the network structure and to fix the location of each directory entry (note that, the data item itself is still not fixed to any node) 48, 58] COMA s ability that data items can automatically migrate or be duplicated looks attractive to researchers in the area of computer architecture [58, 77, 78], however, except for KSR 1 of Kendall Square Research, people in the computer industry have chosen more orthodox CC NUMA architectures. 1.3 Multiprocessor Cache Coherence Problem In modern computer systems, the speed gap between the processors and the memory has been a serious problem. ....

....KSR 1 was a 16 slotted, pipelined, and two levels hierarchical ring. Although KSR 1 might not be commercially successful, its unique design stimulated researchers and many studies have been done such as comparisons of COMA with the cache coherent nonuniform memory access (CC NUMA) architectures [62, 77]. Lam et al. [36] investigated the feasible system size of the hierarchical (two levels) ring multiprocessor for a given bisection bandwidth of the ring. Ravindran and Stumm studied design issues (such as flow control) in two and three levels of rings using mathematical models [55] Scalable ....

X. Zhang and Y. Yan, "Comparative Modeling and Evaluation of CC-NUMA and COMA on Hierarchical Ring Architecture," Transactions on Parallel and Distributed Systems, IEEE, Vol. 6, No. 12, 1316--1331, December 1995.

....slots. Therefore, the transmission delay of a data message on the bidirectional ring is 8tr 1 . We assume the slotted ring configuration for the bidirectional ring. 3 Performance Model The methods that have been used for performance evaluation of computer systems include analytical models ([17]) parametric simulations ( 14] trace driven simulations ( 6] and execution driven simulations ( 5] In this paper, we use the hybrid approach by Barroso and Dubois [1] by extending it to the bidirectional ring and the crossbar. Below we briefly describe the derivation of execution time using ....

X. Zhang and Y. Yan, "Comparative Modeling and Evaluation of CC-NUMA and COMA on Hierarchical Ring Architecture", Transactions on Parallel and Distributed Systems, IEEE, Vol. 6, No. 12, 1316--1331, December 1995.

....a data message is divided into multiple packets and they are sent in (possibly) uncontiguous slots. We assume the slotted ring configuration for the bidirectional ring. 3 Performance Model The methods that have been used for performance evaluation of computer systems include analytical models ([8]) parametric simulations ( 4] trace driven simulations ( 9] and execution driven simulations ( 10] In this paper, we use the hybrid approach by Barroso and Dubois [5] by extending it to the bidirectional ring and the crossbar. Below we briefly describe the derivation of execution time using ....

X. Zhang and Y. Yan, "Comparative Modeling and Evaluation of CC-NUMA and COMA on Hierarchical Ring Architecture ", Transactions on Parallel and Distributed Systems, IEEE, Vol. 6, No. 12, 1316--1331, December 1995.

....a data message is divided into multiple packets and they are sent in (possibly) uncontiguous slots. We assume the slotted ring configuration for the bidirectional ring. 3 Performance Model The methods that have been used for performance evaluation of computer systems include analytical models ([8]) parametric simulations ( 4] trace driven simulations ( 9] and execution driven simulations ( 10] In this paper, we use the hybrid approach by Barroso and Dubois [5] by extending it to the bidirectional ring and the crossbar. Below we briefly describe the derivation of execution time using ....

X. Zhang and Y. Yan, "Comparative Modeling and Evaluation of CC-NUMA and COMA on Hierarchical Ring Architecture", Transactions on Parallel and Distributed Systems, IEEE, Vol. 6, No. 12, 1316--1331, December 1995.

.... architecture (COMA) multiprocessor with hierarchy rings [7] Although KSR 1 might not be commercially successful, its unique design stimulated researchers and many studies have been done such as comparisons of COMA and orthodox cache coherent nonuniform memory access (CC NUMA) architectures ([8, 9]) Lam et al. investigated the feasible system size of the hierarchical (two levels) ring multiprocessor for a given bisection bandwidth of the ring [10] Ravindran and Stumn studied design issues (such as flow control) in two and three levels of ring using mathematical models. Scalable Coherent ....

X. Zhang and Y. Yan, Comparative Modeling and Evaluation of CCNUMA and COMA on Hierarchical Ring Architecture, Transactions on Parallel and Distributed Systems, IEEE, Vol. 6, No. 12, 1316--1331, December 1995.

No context found.

X. Zhang and Y. Yan, "Comparative modeling and evaluation of CC-NUMA and COMA on hierarchical ring architectures", IEEE Transactions on Parallel and Distributed Systems, Vol. 6, No. 12, 1995, pp. 1316-1331.

No context found.

X. Zhang and Y. Yan, " Comparative Modeling and Evaluation of CC-NUMA and CC-COMA on Hierarchical Ring Architectures, " IEEE Transactions on Parallel and Distributed Systems, Vol. 6, No. 12, 1995, pp. 1316--1331.

No context found.

X. Zhang and Y. Yan, "Comparative modeling and evaluation of CC-NUMA and COMA on hierarchical ring architectures", to appear in IEEE Transactions on Parallel and Distributed Systems.

....which could be examined precisely by program execution on a particular architecture. By tracing program executions on the Exemplar, we present overall scaling capability of the architecture. We also compared the system and application performance with the performance of the KSR 1 reported in [13] [14] and [15] The testing programs on the Exemplar were run within the CXpa (CONVEX Performance Analyzer) 2] environment for collecting the execution timing results. The CXpa is a software performance monitor which has the capability to time loops and count cache misses. Although the profiler ....

....a memory module, the memory access to the hot spot is significantly delayed in the KSR 1, but the latency only slightly increases in the Exemplar. ffl Comparing a COMA to a CC NUMA system, if the communication to computation ratio is large, the CC NUMA system is expected to have more advantages [14]. The Barnes application program has low communication to computation ratio which favors the the COMA 18 system. However, the performance was poor when two rings of the KSR 1 were used. We believe this has little to do with the COMA system, but is related to its higher latency of the ....

X. Zhang and Y. Yan, "Comparative modeling and evaluation of CC-NUMA and COMA on hierarchical ring architectures", IEEE Transactions on Parallel and Distributed Systems, Vol. 6, No. 12. 1995, pp. 1316-1331.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC