Processor Hardware Counter Statistics As A First-Class System Resource Xiao Zhang, Sandhya Dwarkadas, Girts Folkmanis, and Kai Shen Department of Computer Science, University of Rochester Today's processors provide a rich source of statistical information on program execution characteristics through hardware counters. However, traditionally, operating system (OS) support for and utilization of the hardware counter statistics has been limited and ad hoc. In this paper, we make the case for direct OS management of hardware counter statistics. First, we show the utility of processor counter statistics in CPU scheduling (for improved performance and fairness) and in online workload modeling, both of which require online continuous statistics (as opposed to ad hoc infrequent uses). Second, we show that simultaneous system and user use of hardware counters is possible via time-division multiplexing. Finally, we highlight potential counter misuses to indicate that the OS should address potential security issues in utilizing processor counter statistics.