Computer producers invest a tremendous measure of energy, assets, and cash in planning new frameworks, and more current designs, and their capacity to diminish costs, charge serious costs, and increase piece of the overall industry relies upon how great these frameworks perform. In this work, we focus on both the framework structure and the compositional plan measures for equal PCs and create strategies to facilitate them We propose a lot of techniques to arrange sellers dependent on assessed focal preparing unit (CPU) execution and foresee CPU execution dependent on equipment parts. For seller characterization, we utilize the most elevated and least assessed execution and recurrence of the events of every merchant in the dataset to make grouping zones. These zones can be utilized to list merchants who produce equipment that fulfills given execution prerequisites. We use multi-layered neural systems for execution expectation, which represents nonlinearity in execution information. A few neural system structures are broken down in contrast with straight, quadratic, and cubic relapse. Examinations show that neural systems can be utilized to get low expectation blunder and the high relationship between anticipated and distributed exhibition esteems, while cubic relapse can deliver higher connection than neural systems when a larger number of information is utilized for preparing than testing. The proposed techniques can be utilized to recognize reasonable equipment substitutions.