C. Alexander, D. Reese, and J. Harden. Near--critical path analysis of program activity graphs. In Proceedings of the 2nd International Workshop on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems, pages 308--317. IEEE Computer Society, Feb. 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... functional units, they can measure whether instructions had sufficient slack to tolerate a slow resource by sending an instruction to a slow resource, and then measuring whether it became critical, according to the critical path predictor from [7] In a different domain, Alexander et al. [1] study the nearcritical paths of the graphs of communication and computation in parallel programs. They propose a Maximum Benefit Metric that quantifies the maximum improvement in runtime possible by reducing the execution time of a section of code. The metric we propose quantifies the maximum ....

C. Alexander, D. Reese, and J. Harden. Near--critical path analysis of program activity graphs. In Proceedings of the 2nd International Workshop on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems, pages 308--317. IEEE Computer Society, Feb. 1994.

....performed on smaller scale systems [10] 2) phase analysis techniques can be employed to partition the optimization process [11] and (3) some useful performance analysis tools are based on complete traces. Near critical path analysis is an example of a tool that relies on unabridged traces [12]. Near critical path analysis includes the synergistic effects of activities on the k longest execution paths to quantify the overall benefit associated with improving specific critical path activities. Scalability is addressed using a distributed design that replicates the SPIscope for clusters ....

C. Alexander, D. Reese, and J. Harden, "Near--Critical Path Analysis of Program Activity Graphs," Proc. IEEE MASCOTS`94, Jan. 31 -- Feb. 2, 1994.

....begins at the root vertices and diffuses to all descendant vertices. In the synchronous variation, a vertex will not diffuse a computation to its descendants until all incoming computations are received. A version of the synchronous algorithm with adaptations to accommodate multigraphs is given in [4]. C. Critical Path Method The critical path method is an operational research algorithm for finding the longest path(s) through an activity on edge network [5] The critical path method calculates early start and early finish times for each activity in a forward pass through the network. Late ....

....longest paths. For a multigraph containing n vertices and e edges, the worst case time and memory requirements of the algorithm are in O(kne) and O(kn 2 e) respectively. A more straightforward approach is to simply extend the longest path algorithm to find the k longest paths as described in [4]. Since the extended algorithm maintains an array of k (fixed size) path description records for each vertex, and a descriptor is required to represent each edge, the storage requirements are in O(kn e) The worst case time complexity of the algorithm is in O(ke) B. Branch and Bound ....

[Article contains additional citation context not shown here]

C. A. Alexander, D. S. Reese, and J. C. Harden, "Near--critical path analysis of program activity graphs," in Proc. 2nd Int. Workshop on Modeling, Analysis, and Simulation of Comput. and Telecommun. Syst., IEEE Comput. Soc., Jan. 31--Feb. 2, 1994, pp. 308--317.

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