Measuring the energy consumption of software components is a major building block for generating models that allow for energy-aware scheduling, accounting and budgeting. Current measurement techniques focus on coarse-grained measurements of application or system events. However, fine grain adjustments in particular in the operating-system kernel and in application-level servers require power profiles at the level of a single software function. Until recently, this appeared to be impossible due to the lacking fine grain resolution and high costs of measurement equipment. In this paper we report on our experience in using the Running Average Power Limit (RAPL) energy sensors available in recent Intel CPUs for measuring energy consumption of short code paths. We investigate the granularity at which RAPL measurements can be performed and discuss practical obstacles that occur when performing these measurements on complex modern CPUs. Furthermore, we demonstrate how to use the RAPL infrastructure to characterize the energy costs for decoding video slices.

Measuring the energy consumption of software components is a major building block for generating models that allow for energy-aware scheduling, accounting and budgeting. Current measurement techniques focus on coarse-grained measurements of application or system events. However, fine grain adjustments in particular in the operating-system kernel and in application-level servers require power profiles at the level of a single software function. Until recently, this appeared to be impossible due to the lacking fine grain resolution and high costs of measurement equipment. In this paper we report on our experience in using the Running Average Power Limit (RAPL) energy sensors available in recent Intel CPUs for measuring energy consumption of short code paths. We investigate the granularity at which RAPL measurements can be performed and discuss practical obstacles that occur when performing these measurements on complex modern CPUs. Furthermore, we demonstrate how to use the RAPL infrastructure to characterize the energy costs for decoding video slices.

Massive data centers housing thousands of computing nodes have become commonplace in enterprise computing, and the power consumption of such data centers is growing at an unprecedented rate. Exacerbating such costs, data centers are often over-provisioned to avoid costly outages associated with the potential overloading of electrical circuitry. However, such over provisioning is often unnecessary since a data center rarely operates at its maximum capacity. It is imperative that we realize effective strategies to control the power consumption of the server and improve the energy efficiency of data centers. Adding to the problem is the inability of the servers to exhibit energy proportionality which diminishes the overall energy efficiency of the data center. Therefore in this dissertation, we investigate whether it is possible to achieve energy proportionality at the server- and cluster-level by efficient power and resource provisioning. Towards this end, we provide a thorough analysis of energy proportionality at the server and cluster-level and provide insight into the power saving opportunity and mechanisms to improve energy proportionality. Specifically, we make the following contribution at the server-level using enterprise-class workloads. We analyze the average power consumption of the full system as well as the subsystems and describe the energy proportionality of these components, characterize the instantaneous power profile of enterprise-class workloads using the on-chip energy meters, design a runtime system based on a load prediction model and an optimization framework to set the appropriate power constraints to meet specific performance targets and then present the effects of our runtime system on energy proportionality, average power, performance and instantaneous power consumption of enterprise applications. We then make the following contributions at the cluster-level. Using data serving, web searching and data caching as our representative workloads, we first analyze the component-level power distribution on a cluster. Second, we characterize how these workloads utilize the cluster. Third, we analyze the potential of power provisioning techniques (i.e., active low-power, turbo and idle low-power modes) to improve the energy proportionality. We then describe the ability of active low-power modes to provide trade-offs in power and latency. Finally, we compare and contrast power provisioning and resource provisioning techniques. This thesis sheds light on mechanisms to tune the power provisioned for a system under strict performance targets and opportunities to improve energy proportionality and instantaneous power consumption via efficient power and resource provisioning at the server- and cluster-level. / Ph. D.