Parallel principles are the most effective way how to increase parallel computer performance and parallel algorithms (PA) too. In this sense the paper is devoted to a complex performance evaluation of chosen PA. At first the paper describes very shortly PA and then it summarized basic concepts for performance evaluation of PA. To illustrate the analyzed evaluation concepts the paper considers in its experimental part the results for real analyzed examples of discrete fast Fourier transform (DFFT). These illustration examples we have chosen first due to its wide application in scientific and engineering fields and second from its representation of similar group of PA. The basic form of parallel DFFT is the one-dimensional (1-D), unordered, radix-2 algorithm which uses divide and conquer strategy for its parallel computation. Effective PA of DFFT tends to computing one-dimensional FFT with radix greater than two and computing multidimensional FFT by using the polynomial transfer methods. In general radix-q DFFT is computed by splitting the input sequence of size s into q sequences each of them in size n/q, computing faster their q smaller DFFT's, and then combining the results. So we do it for actually dominant asynchronous parallel computers based on Network of workstations (NOW) and Grid systems.