Abstract: One example embodiment is a method that simulates a sampling period of an application to collect execution counts of basic blocks and compute cycles per instruction (CPI) data. A non-sampling period of the application is simulated to collect execution counts of basic blocks, and a comparison of the execution counts collected during the sampling period is performed to the execution counts collected during the non-sampling period. Based on the comparison, a determination is made whether to estimate CPI for the basic blocks during the non-sampling period using the CPI data collected during the sampling period.

Abstract: Nodes interconnected by a network have their substantially parallel execution simulated. Substantially parallel execution of the nodes during a current quantum of simulation time having a quantum length is simulated. Simulation of execution can result in simulation of inter-node data packets being transmitted over the network. When the current quantum of simulation time has elapsed, simulation of execution of the nodes is synchronized. If no inter-node data packets were transmitted in simulation during the current quantum of simulation time, then the quantum length is increased. If one or more inter-node data packets were transmitted in simulation during the current quantum of simulation time, then the quantum length is decreased. This process is then repeated for a next quantum of simulation time having the quantum length as has been increased or decreased.

Abstract: One example embodiment is a method that simulates a sampling period of an application to collect execution counts of basic blocks and compute cycles per instruction (CPI) data. A non-sampling period of the application is simulated to collect execution counts of basic blocks, and a comparison of the execution counts collected during the sampling period is performed to the execution counts collected during the non-sampling period. Based on the comparison, a determination is made whether to estimate CPI for the basic blocks during the non-sampling period using the CPI data collected during the sampling period.

Abstract: A hybrid prophet/critic predictor includes a first branch predictor to provide a first branch prediction for a branch under prediction (BUP) based on a branch history of the BUP and/or a program counter, and also includes a second branch predictor to provide a second branch prediction for the BUP based on a branch future of the BUP.