McCreary, C.L., "Partitioning for Parallelization Using Graph Parsing", Technical Report CSE-TR-91-17, Auburn University, 1991.  Home/Search   Document Details and Download   Summary   Related Articles

....on the concept of a clan, which originated in the work of Ehrenfeucht and Rozenberg [1] A clan is defined informally as a subset of the vertices of a graph such that every element not in the subset is related in the same way (i.e. ancestor, descendent, or neither) to each member in the subset. [6] Clans are identified as trivial, independent, linear, or primitive, where each single node is a trivial clan, and primitive clans may be decomposed into independent and linear clans. Nodes within an independent clan could be executed in parallel, whereas those within a linear clan have ....

....multi stage optimization algorithm [4] The shortest path represents the sequence of decisions which will result in the lowest total cost for executing the linear clan. IV. Processor Assignment When the decision for an independent child is to parallelize, processor assignment must be considered [3, 6]. For example, if the independent child consists of two nodes, each of which communicates to a third node, the third node should be placed on the same processor as the previous with the more expensive communication. Similarly, if communication times are constant, and two nodes communicate to a ....

McCreary, C.L., "Partitioning for Parallelization Using Graph Parsing", Technical Report CSE-TR-91-17, Auburn University, 1991.

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