| P-Z. Lee and W.Y. Chen, "Compiler Techniques for Determining Data Distribution and Generating Communication Sets on Distributed -Memory Multicomputers," Proc. 29th Hawaii Int'l. Conf. System Sciences, vol. 1, pp. 537-546, Jan. 1996. |
....index sets identification. By using their method, closed form expressions for index sets of arrays that were mapped to processors using one level mapping can be translated to closed form expressions for index sets of arrays that were mapped to processors using two level mapping and vice versa. In [17], a similar approach that addressed the problems of the index sets and the communication sets identification for array statements with BLOCK CYCLIC(c) distribution was presented. Lee and Chen [17] derived communication sets for statements of arrays that were distributed in arbitrary ....
....for index sets of arrays that were mapped to processors using two level mapping and vice versa. In [17] a similar approach that addressed the problems of the index sets and the communication sets identification for array statements with BLOCK CYCLIC(c) distribution was presented. Lee and Chen [17] derived communication sets for statements of arrays that were distributed in arbitrary BLOCK CYCLIC(c) fashion. They also presented closed form expressions of communication sets for restricted block sizes. In [9] an approach for generating communication sets by computing the intersections of ....
P-Z. Lee and W.Y. Chen, "Compiler Techniques for Determining Data Distribution and Generating Communication Sets on Distributed -Memory Multicomputers," Proc. 29th Hawaii Int'l. Conf. System Sciences, vol. 1, pp. 537-546, Jan. 1996.
....execute each set of iterations in different PEs independently. Figure 1 shows two programs: a) a sequential program for solving a linear system AX = B, and (b) its corresponding doall loop program. The corresponding SPMD program in which data arrays are distributed by cyclic(b) can be found in [10]. Readers can find that there is a one to one correspondence between statements in the original sequential program (which have been rewritten after performing loop transformations) and its corresponding doall version program. For this reason, without any confusion, in the sequel we will frequently ....
....execution time on the nCUBE 2 computer. We now continue our sample example of the linear system. Let the time of executing the LU decomposition be TLU ; the time of executing two triangular linear systems be T 2TLS ; and the total execution time be T . Then, from the corresponding SPMD program in [10], we can formulate T , TLU , and T 2TLS as follows. T = TLU T 2TLS , TLU = P m= N b) i 1 =1 P N i 2 =1 Phi 9 t f P b i 3 =1 P b i 4 =i3 1 Gamma t f (m Gamma ( i 1 Gamma 1) N b (i 2 Gamma 1) b i 3 ) 1) t x Delta P m= N b) i 5 =i1 P b i 6 =1 P b i 7 =1 ....
[Article contains additional citation context not shown here]
P.-Z. Lee and W. Y. Chen. Compiler techniques for determining data distribution and generating communication sets on distributed-memory multicomputers. Technical Report TR-IIS-95-007, Institute of InformationScience, Academia Sinica, 1995.
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