Evaluating thread level parallelism based on optimum cache architecture | IEEE Conference Publication | IEEE Xplore