F. Desprez, J. Dongarra, A. Petitet, C. Randriamaro, and Y. Robert. Scheduling blockcyclic array redistribution. IEEE TPDS, 9(2):192-205, 1998. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Block-Cyclic Redistribution over Heterogeneous Networks - Bhat, Prasanna, Raghavendra (1998) (3 citations) (Correct)
....performs the x (K; P ) communication in K communication steps. The indirect approach performs x (K; P ) in atmost dlog 2 Ke 2 steps. Here, the array elements are communicated in a combine and forward manner. The hybrid approach is a combination of the direct and indirect approaches. In [2], communication schedules are developed for the general redistribution problem of cyclic(r) over a set of P processors, to cyclic(s) over a different set of Q processors. Graph matching algorithms are used to develop communication schedules in this work. These techniques have two important ....
F. Desprez, J. Dongarra, A. Petitet, C. Randriamaro, and Y. Robert. Scheduling block-cyclic array redistribution. IEEE Trans. Parallel and Distributed Systems, 9(2):192-- 205, Feb. 1998.
Sparse Matrix Block-Cyclic Redistribution - Bandera, Zapata (1999) (Correct)
....Lim et The work described in this paper was supported by the Ministry of Education and Science (CICYT) of Spain under project TIC96 1125 C03 and the EU Brite Euram project BE 1564. al. 7] study the redistribution of a dense vector based on the notion of the circulant matrix, while in [4] an optimization of the message scheduling using coloring graphs is presented. When sparse matrices are used different data structures and distributions can be used in order to improve the performance. In this paper, we present an analysis of the redistribution algorithm for sparse matrices ....
F. Desprez, J. Dongarra, C. Randriamaro, Y. Robert, Scheduling Block-Cyclic Array Redistribution, IEEE Transactions on Parallel and Distributed Systems, Vol. 9, No. 2, February 1998.
Efficient Algorithms for Block-Cyclic Array Redistribution.. - Park, Prasanna, al. (Correct)
....same . Node contention occurs . Does not minimize the transmission cost Caterpillar [11] Simple scheduling algorithm . Index computation by scanning the array segments . No node contention . Does not minimize the transmission cost and the number of communication steps Bipartite Matching Scheme [5] . Large schedule computation overhead Schedule computation time: O( P Q) 4 ) No node contention . Stepwise strategy: minimizes the number of communication steps . Greedy strategy: minimizes the transmission cost Our Scheme . Fast schedule and index computations Schedule computation time: ....
....cost Table 1: Comparison of various schemes for array redistribution. However, this algorithm does not fully utilize the network bandwidth i.e. the size of the data sent by the nodes in a communication step varies from node to node. This leads to increased data transfer cost. The schemes in [5] reduce the data transfer cost, however, the schedule computation cost is significant. The bipartite graph matching used in [5] takes O( P Q) 4 ) time. On a stateof the art workstation, this time is in the range of 100 s of msecs for P and Q of interest. For problems of interest, the schedule ....
[Article contains additional citation context not shown here]
F. Desprez, J. Dongarra, A. Petitet, C. Randriamaro and Y. Robert. Scheduling Block-Cyclic Array Redistribution. University of Tennessee Computer Science Technical Report, UT-CS97 -349, LAPACK Working Note 120, February 1997.
Optimal Task Scheduling to Minimize Inter-Tile Latencies - Rastello, Rao, Pande (1998) (Correct)
....using unimodular transformations. However, this work is not suited to DSMs as it involves complex data partitioning and assignment of iterations to processors. Tiling is more attractive in such situations where data and iteration space partitioning can be kept simple and efficient. Many approaches [1, 3, 8, 10] use the assumption of atomic tiles to derive the optimal tile size and shape. These approaches involve re distribution of data among processors in order to improve data locality thereby addressing the issue of memory to tile latencies. On the other hand, we perform loop tiling and attempt to ....
F. Desprez et al. Scheduling Block-Cyclic Array Redistribution. In Parallel Computing '97 (ParCo97). North-Holland, September 1997.
From Heterogeneous Task Scheduling to Heterogeneous Mixed Data.. - Suter, al. (2003) Self-citation (Desprez) (Correct)
No context found.
Frdric Desprez, Jack Dongarra, Antoine Petitet, Cyril Randriamaro, and Yves Robert. Scheduling Block-Cyclic Array Redistribution. IEEE Transactions on Parallel and Distributed Systems, 9(2):192205, February 1998. 14 F. Suter et al.
Data Redistribution Algorithms for Heterogeneous Processor.. - Renard, Robert, Vivien (2004) (Correct)
No context found.
F. Desprez, J. Dongarra, A. Petitet, C. Randriamaro, and Y. Robert. Scheduling blockcyclic array redistribution. IEEE TPDS, 9(2):192-205, 1998.
Data Redistribution Algorithms For Heterogeneous Processor.. - Renard, Robert, Vivien (2004) (Correct)
No context found.
F. Desprez, J. Dongarra, A. Petitet, C. Randriamaro, and Y. Robert. Scheduling block-cyclic array redistribution. IEEE Trans. Parallel Distributed Systems, 9(2):192--205, 1998.
Data Redistribution Algorithms For Heterogeneous Processor.. - Renard, Robert, Vivien (2004) (Correct)
No context found.
F. Desprez, J. Dongarra, A. Petitet, C. Randriamaro, and Y. Robert. Scheduling blockcyclic array redistribution. IEEE Trans. Parallel Distributed Systems, 9(2):192--205, 1998.
Unknown - De Recherche (Correct)
No context found.
F. Desprez, J. Dongarra, A. Petitet, C. Randriamaro, and Y. Robert. Scheduling Block-Cyclic Array Redistribution. IEEE Transaction on Parallel and Distributed Systems, 9(2):192205, 1998.
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