Since Mars rovers have limited life span, NASA wants to maximize the exploration activities during this period. Rock sample analysis is one of the main tasks of rover missions. Traditionally, rock selection is decided by human operators. Due to long communication delay, manual selection process is time-consuming. There is a strong need to develop an automatic software system to automate the process. We propose a novel and high performance approach to enhancing rock selection process. We explicitly take advantage of the availability of LIBS instrument in the new generation of Mars rover. First, we use LIBS to quickly sample the neighborhood of the rover. LIBS can collect samples in seconds. Our software algorithms can quickly analyze the LIBS data and determine whether there are any interesting chemical elements. If yes, the APXS instrument will be activated. Otherwise, the rover will move to a new location and start the process again. In Phase I, we have demonstrated that our smart processing tools using actual Mars data and our results are more consistent than a current method. Moreover, our tools can implemented in a parallel processing system to achieve real-time performance. Our parallel processing system utilizes multi-core CPUs for distributed processing and we have used such processing architecture for speech and genomic processing.