PowerSched - Managing Power Consumption in Overprovisioned Systems

This paper focuses on applying Q learning techniques in a home energy management agent where the agent learns to find the optimal sequence of turning off appliances so that the appliances with higher priority will not be switched off during peak demand period or power consumption management. The policy based home energy management determines the optimal policy at every instant dynamically by learning through the interaction with the environment using one of the reinforcement learning approaches called Q-learning. The Q-learning home power consumption problem formulation consisting of state space, actions and reward function is presented in this paper. The simulation results show that the proposed Q-learning based power consumption management is very effective and enables the users to have minimum discomfort during participation in peak demand management or at the time when power consumption management is essential when the available power is rationale.

Purpose. Improving the principles of management of photovoltaic system with storage battery and with autonomous functioning during daylight hours for a local object, connected to the grid with multi-zone payment when excluding the generation of energy into the grid. Methodology. Modeling and analysis of energy processes in the photovoltaic system was performed using the Matlab software package. The simulation model of energy processes is based on calculated expressions taking into account the characteristics of the battery. Operability of the proposed solutions are confirmed on an experimental setup based on a standard hybrid inverter. Results. It’s shown, that due to the battery energy during the most loaded peak hours and part of the daytime the system operates autonomously and does not depend on possible violations of the quality of electricity in the grid. Scenarios of the recommended load schedule are proposed in accordance with the ratio of the predicted value of the daily energy generation of the photovoltaic battery to its possible maximum value. A simulation model of energy processes in the system with the correction of the recommended load value was developed. Originality. A method of the recommended load calculation with current correction for the actual generation and degree of battery charge is proposed, which allows taking into account differences the actual generation of the photovoltaic battery from its predicted value and the actual load from the recommended one. Practical value. The obtained solutions are the basis for the design of new and modernization of existing photovoltaic systems of local objects using software and hardware complexes for power consumption management.

As technology scales for increased circuit density and performance, the management of power consumption in embedded systems is becoming critical. Because the operating system (OS) is a basic component of the embedded system, the reduction and characterization of its energy consumption is a main challenge for the designers. In this work, a flow of low power OS energy characterization is introduced. The variation of the energy and power consumption of the embedded OS services is studied. The remainder of this article details the methods used to determine energy and power overheads of a set of basic services of the embedded OS: scheduling, context switch and inter-process communication. The impacts of hardware and software parameters like processor frequency and scheduling policy on the energy consumption are analyzed. Also, models and laws of the power and energy are extracted. Then, to quantify the low power OS energetic overhead, the obtained models are integrated in the system level design. Our method allows estimating the energy consumption of the low power OS services when running an application on a specific hardware platform.

Wireless networking is a crucial component of communication in this era of digitization. The key characteristics that determine a wireless network's performance are smooth connectivity, quicker communication, and increased customer satisfaction. Scalability deals with the erratic workload and new resources that are added to the network. Scalability is crucial for both Quality of Experience (QoE) and Quality of Service (QoS) in such a power-constrained environment because wireless devices are battery-operated. Managing power consumption and network performance simultaneously in scalable networking scenario is quite challenging, especially during handover. This research proposed a methodology to assess the impact of scalability on wireless node power consumption and network performance in order to develop energy-efficient handoff mechanisms. The Type-I and Type-II wireless node sets were used between the WiFI and WiMAX networks in a vertical handover scenario. In the study, two distinct scalability scenarios were taken into account. Three distinct voltage levels—cutoff voltage, nominal voltage, and charge voltage —had been used to examine the consumption of power. Throughput, residual energy, and Packet Delivery Ratio (PDR) were computed to examine the network's performance in a scalable, power-constrained context. Two different scalability scenarios were considered in the study. The power consumption was investigated using three different voltage levels: charge voltage, nominal voltage, and cutoff voltage. The network's performance under scalable, power-constrained conditions was investigated by computing throughput, packet delivery ratio (PDR), and residual energy. The result obtained in the study showed the variation in battery drainage of wireless nodes under the effect of various scalability factors. This study will provide a blueprint for developing novel algorithms which eventually will improvise network performance and strengthen the base of Green Networking.

Traditional multilevel SRAM-based cache hierarchies, especially in the context of chip multiprocessors (CMPs), present many challenges in area requirements, core--to--cache balance, power consumption, and design complexity. New advancements in technology enable caches to be built from other technologies, such as Embedded DRAM (EDRAM), Magnetic RAM (MRAM), and Phase-change RAM (PRAM), in both 2D chips or 3D stacked chips. Caches fabricated in these technologies offer dramatically different power-performance characteristics when compared with SRAM-based caches, particularly in the areas of access latency, cell density, and overall power consumption. In this article, we propose to take advantage of the best characteristics that each technology has to offer through the use of Hybrid Cache Architecture (HCA) designs. We discuss and evaluate two types of hybrid cache architectures: intercache-Level HCA (LHCA), in which the levels in a cache hierarchy can be made of disparate memory technologies; and intracache-level or cache-Region-based HCA (RHCA), where a single level of cache can be partitioned into multiple regions, each of a different memory technology. We have studied a number of different HCA architectures and explored the potential of hardware support for intracache data movement and power consumption management within HCA caches. Utilizing a full-system simulator that has been validated against real hardware, we demonstrate that an LHCA design can provide a geometric mean 6% IPC improvement over a baseline 3-level SRAM cache design under the same area constraint across a collection of 30 workloads. A more aggressive RHCA-based design provides 10% IPC improvement over the baseline. A 2-layer 3D cache stack (3DHCA) of high density memory technology within the same chip footprint gives 16% IPC improvement over the baseline. We also achieve up to a 72% reduction in power consumption over a baseline SRAM-only design. Energy-delay and thermal evaluation for 3DHCA are also presented. In addition to the fast-slow region based RHCA, we further evaluate read-write region based RHCA designs.

As HPC systems increase in scale and capability, the cost of supplying power to these systems grows significantly. This introduces the urgent need for energy efficient computing through the management of power consumption. The PowerStack initiative defines a holistic power management framework at three levels, i.e., at the cluster level, the job level, and the node level [1] –[3]. It is expected that a cluster will be given a system-wide power budget (aka an allocated power budget), which can be intelligently managed and distributed at the job and node levels. Our work provides a method for intelligently managing node-level power during execution.

Virtualization is the main building block of many architectures and systems from embedded computing to large scale data-centers. Managing efficiently computing resources and their power consumption becomes fundamental to optimize the performance of the workloads running on those systems, however, hardware tools like Intel RAPL can only introduce power caps without considering performance. This paper presents a performance-aware power capping orchestrator for the Xen hypervisor. The tool exploits hybrid power management techniques to minimize power consumption respecting a given SLA, leveraging RAPL and overcoming its limitations. Experimental evaluation shows that the proposed approach guarantees good results for almost all the analyzed benchmarks (e.g. CPU-, memory- and IO-bound).

Wireless multimedia sensor network (WMSN) is one of broad wide application for developing a smart city. Each node in the WMSN has some primary components: sensor, microcontroller, wireless radio, and battery. The components of WMSN are used for sensing, computing, communicating between nodes, and flexibility of placement. However, the WMSN technology has some weakness, i.e. enormous power consumption when sending a media with a large size such as image, audio, and video files. Research had been conducted to reduce power consumption, such as file compression or power consumption management, in the process of sending data. We propose Green Communication (GeCom), which combines power control management and file compression methods to reduce the energy consumption. The power control management method controls data transmission. If the current data has high similarity with the previous one, then the data will not be sent. The compression method compresses massive data such as images before sending the data. We used the low energy image compression algorithm algorithm to compress the data for its ability to maintain the quality of images while producing a significant compression ratio. This method successfully reduced energy usage by 2% to 17% for each data.

The 21st century is grappling with numerous challenges, with sustainability being a central concern. While innovation has propelled humanity into the modern era, it has also contributed to environmental issues, particularly through the extensive use of fossil fuels. In India, a developing nation with limited power generation capacity, the rapid shift toward electric vehicles (EVs)—driven by rising fuel prices—could lead to a substantial increase in electricity consumption. This shift places a significant strain on the power grid, making it essential to understand the impact on power generation and distribution systems. The study compares traditional charging models with smart charging solutions, finding that while the current impact is relatively minor, smart charging is expected to become increasingly effective by 2034 and 2040, when EV adoption reaches critical thresholds. The smart charging model proves to be more efficient in managing power consumption and can help optimize the power distribution system, providing a viable solution for the anticipated surge in demand

This study proposes a coordinated multi-objective optimization framework for scheduling and managing power consumption in green power systems with diverse energy devices. The system incorporates renewable sources, energy storage units, and controllable loads to enhance energy efficiency, reduce costs, and minimize environmental impact. A multi-objective evolutionary algorithm is employed to balance conflicting goals such as operational cost, emission reduction, and user comfort. Device-level coordination, including electric vehicles, HVAC systems, and smart appliances, is integrated to enable demand-side flexibility. Simulation results demonstrate the effectiveness of the proposed strategy in achieving optimal scheduling under dynamic load and generation conditions, contributing to the realization of intelligent, low-carbon energy systems.

In this paper, we present a new methodology for managing power consumption of networks-on-chips (NOCs). A power management problem is formulated for the first time using closed-loop control concepts. We introduce an estimator and a controller that implement our power management methodology. The estimator is capable of very fast and accurate tracking of changes in the system parameters. Parameters estimated are used to form the system model. Our system model combines node and network centric power management decisions. Node centric power management assumes no a priori knowledge of requests coming in from outside the core. Thus, it implements a more traditional dynamic voltage scaling and power management control algorithms. Network-centric power management utilizes interaction with the other system cores regarding the power and the quality of service (QoS) needs. The overall system model is based on Renewal theory and, thus, guarantees globally optimal results. We introduce a fast optimization method that runs multiple orders of magnitude faster than the previous optimization approaches while still having the same accuracy in obtaining the power management control. Finally, our controller implements the results of optimization in either hardware or software. The new methodology for power management of NOCs is tested on a system consisting of four satellite units, each implementing an estimator and a controller capable of both node and network centric power management. Our results show large savings in power with good QoS.

This is a concise review of the recent work on devices, circuits and architectures for possible hybrid semiconductor/nanodevice integrated circuits based on nanowire crossbars, with similar, simple, two-terminal devices formed at each crosspoint. Special attention is given to the so-called CMOL' variety of the hybrids, in which the crossbar is connected to the underlying CMOS circuit, using an area-distributed, pin-based interface. Recent detailed studies have shown that digital hybrid circuits may extend the Moore's Law progress of integrated circuits for 10 to 15 years. Even more impressive, mixed-signal neuromorphic networks (CrossNets) may provide unparalleled performance for some important information processing tasks, and in future may become the first hardware basis for challenging the mammal cerebral cortex in density, far exceeding it in speed, at manageable power consumption. Recently, the hybrid circuit concept received a strong boost from the experimental demonstrations of reproducible crosspoint devices (latching switches) based on amorphous silicon and metal oxides, and of nanowire crossbars with 15-nm-scale half-pitch. However, CMOL technology still faces several significant challenges, briefly discussed in the last section.

Platforms with multiple cores are now prevalent everywhere from desktops and graphics processors to laptops and embedded systems. By adding more parallel computational resources while managing power consumption, multicore platforms offer better programmability, performance, and power efficiency. Signal processing systems of tomorrow will be and must be implemented on platforms with multiple cores. Writing efficient parallel applications that utilize the computing capability of many processing cores require some effort. Signal processing algorithm designers must understand the nuances of a multicore computing engine; only then can the tremendous computing power that such platforms provide be harnessed efficiently. To give a thorough perspective of the area, we have organized two special issues on this topic.

In this paper, we design a novel two-channel sphygmus monitoring system, which can synchronously measure the pulse signals and their difference one from radial artery and ulnar artery. This system uses the CM-01B vibration sensor as a sphygmus sensor, utilizes the WiFi module to transmit data, makes use of the PC software to process and display real-time data of sphygmus. The pint-sized and portable sphygmus monitoring bracelet consists of a single chip microcomputer and its peripheral circuit, WiFi communication module, LCD, sphygmus monitoring algorithm, status judgment algorithm and software filtering algorithm. We have performed a great number of experiments on power consumption management, data transmission, interference resistance, and put forward the corresponding hardware or software solutions. Current results demonstrate that our design is effective to monitor the human sphygmus, and is also of high importance to the health caring for the elder people.

Foreign experience in applying the principles of power consumption management is analyzed in the paper. A comparative analysis of the main indicators of the load schedules of the power systems of Azerbaijan, Turkey, Ukraine, Russia and Jordan was carried out and it was found out that there is significant potential for improving such indicators of the load schedules as variation factor and the fill factor for the Azerbaijan power system. The positive effects of equalization of the load schedules of the power system of Azerbaijan are studied, based on the technical and economic indicators of the main generating capacities.