S. Sahni and V. Thanvantri, "Parallel computing: Performance metrics and models." Research Report, Computer Science Department, University of Florida, May 1995.  Home/Search   Document Details and Download   Summary   Related Articles

....this paper, problem size refers to the work to be performed and we will use problem size and work alternatively. on how the performance power is defined and measured, different scalability metrics have been proposed and have been used regularly in evaluating parallel algorithms and architectures [5, 6, 7, 8, 9]. Execution time is the ultimate measure of interest in practice. Scalability, however, has been traditionally studied separately. Its relation to time has not been revealed. For this reason, though scalability measurement has been used extensively in the parallel processing community, some ....

....an algorithm machine combination in maintaining uniprocessor utilization. Scalability is a distinct metric for scalable computing. While scalability has been widely used as an important property in analyzing algorithms and architectures, execution time is the dominant metric of parallel processing [9]. Scalability study would have little practical impact if it could not provide useful information on time variation in a scalable computing environment. The relation between scalability and execution time is revealed in this study. Experimental and theoretical results show scalability is a good ....

S. Sahni and V. Thanvantri, "Parallel computing: Performance metrics and models." Research Report, Computer Science Department, University of Florida, May 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