An efficient search algorithm for cache simulation platform

In order to reduce the design risk, cache designers need to evaluate various solutions under cache simulation platform to determine the optimal one. It will consume lots of time to exhaustive all of solutions for benchmarks, since the simulation platform is generally inefficient. In this paper, an efficient search algorithm is proposed to find the optimal cache solution efficiently. It simplifies the search complexity by reducing the search times and still guarantees the accuracy by the predictive technology for best search space and search order of cache parameters. In our experiments, with twice iterations, the efficient search algorithm will save about 68% of the simulation times with almost 75% accuracy.