Abstract
AbstractThe onset of Deep Sub Micron (DSM) technology has driven the computing world towards billion device multi-GHz processor architectures leading to a stiff upward curve in power consumption and power density (W/cm 2). In this paper we develop a graph-based power model for multiprocessors that predicts power requirements across the components of the cluster (Compute node, Memory and Network system) at various hierarchical levels when applications are run. PASCOM proposes new metrics for power measurement that integrates execution module characteristics with power dissipation metrics. The PASCOM model is applied to Memory In Processor chip and we study power consumption for parallel scientific applications from SPLASH2 and NAS Parallel suite. Total power dissipated varies by 15%. However, the static and dynamic power dissipation exhibit up to 33% and 60% variation respectively due to workload characteristics.KeywordsPower DissipationPower ModelPower EstimationExecution ModuleInstruction ExecutionThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
