In Software Engineering simulators are typically used for Software Performance En- gineering (SPE). It is important that the simulations are accurate in order to allow engineers to predict the performance in detail. Palladio is one of these approaches. Currently, Palladio only supports single-core CPU simulators, but there is also an auxiliary approach for multicore simulation. The main problem of this approach is the huge inaccuracy, which is about 74% with 16 cores. This bachelor thesis aims to investigate and improve Palladio’s performance in hardware CPU simulation and performance prediction. This work presents a new approach for connecting a multicore CPU simulator to Palladio to improve the simulation accuracy. The result of this thesis is a conceptual implemen- tation of an embedded multicore CPU Simulator in Palladio to enable more accurate multicore performance predictions. The presented approach enables Palladio to connect to a multicore simulator called MaxSim via a Java prototype, but the predictions aren’t more accurate in general. With a mean speedup deviation of 67.81% at 16 cores, the simulation is only slightly more accurate for the tested system.