After the parallel algorithm has been written as well as executed performance of both algorithms is one of other main concerns. In order to examine performance of algorithms there are numerous types of performance measurement tools. Measurements tools depend not only on parallel algorithm however need to collect data from operating system and hardware being used in order to offer effective utilisation of tools. For a given parallel computer as number of processors and its computational power raises the complexity and volume of data for performance analysis considerably increases. The collected data for measurement is always so difficult for tools to store and process it. The task of measuring performance of parallel programs has been splitted in two parts.
1) Performance Analysis: It gives the important information to programmers from large chunk of statistics available of program whilst in execution mode or from output data.
2) Performance Instrumentation: Its emphasis on how to resourcefully collect information about computation of parallel computer.