Résumé : Predictive power management provides reduced power consumption and increased performance compared to reactive schemes. It effectively reduces the lag between workload phase changes and changes in power adaptations since adaptations can be applied immediately before a program phase change. To this end we present the first analysis of prediction for power management under SYSMark2007. Compared to traditional scientific/computing benchmarks, this workload demonstrates more complex core active and idle behavior. We analyze a table based predictor on a quad-core processor. We present an accurate run-time power model that accounts for fine-grain temperature and voltage variation. By predictively borrowing power from cores, our approach provides an average speedup of 7.3% in SYSMark2007.