Metrics For Performance Evaluation
In this section, we would highlight different kinds of metrics involved for analyzing the performance of parallel algorithms for parallel computers.
The running time is the amount of time consumed in implementation of an algorithm for a given input on the N-processor based parallel computer. The running time is indicated by T(n) where n shows the number of processors employed. If the value of n is equal to 1, then the case is same to a sequential computer. The relation among Execution time vs. Number of processors is shown in Figure 1.
Execution Time vs. number of processors
It can be simply seen from the graph that as the number of processors increases, firstly the implementation time reduces but after a certain optimum level the implementation time increases as number of processors increases. This inconsistency is because of the overheads involved in increasing the number of processes.