CPU Frequency Scaling Research Articles

CPU Frequency Scaling Optimization in Sustainable Edge Computing

Sustainable edge computing (SEC) is a promising technology that can reduce energy consumption and computing latency for the mobile Internet of things (IoT). By collecting renewable energy such as solar or wind energy from the environment, a sustainable cloudlet outside the electric grid can provide powerful computing capabilities for resource-constrained mobile IoT devices. In the real world, the density of sustainable energy can vary significantly over time. Therefore, the SEC cloudlet needs to dynamically adjust the clock frequency to balance energy consumption and computing latency. In this paper, we consider the limited energy storage of the cloudlet and the dynamic intensity of renewable energy, and then develop offline optimal CPU frequency scaling policies that (a) maximize the computing power of the cloudlet within a certain period of time, and (b) minimize the execution time given tasks offloaded to the cloudlet. An optimal tightest string policy is proposed to solve the optimization problem. In addition, a dynamic programming (DP) based suboptimal solution is introduced to simplify the practical implementation. How to design an online CPU frequency management strategy is also briefly discussed.

Correlation of Performance Optimizations and Energy Consumption for Stencil-Based Application on Intel Xeon Scalable Processors

This article provides a comprehensive study of the impact of performance optimizations on the energy efficiency of a real-world CFD application called MPDATA, as well as an insightful analysis of performance-energy interaction of these optimizations with the underlying hardware that represents the first generation of Intel Xeon Scalable processors. Considering the MPDATA iterative application as a use case, we explore the fundamentals of energy and performance analysis for a memory-bound application when exposed to a set of optimization steps that increase the application performance, by improving the operational intensity of code and utilizing resources more efficiently. It is shown that for memory-bound applications, optimizing toward high performance could be a powerful strategy for improving the energy efficiency as well. In fact, for the considered performance optimizations, the energy gain is correlated with the performance gain but with varying degrees. As a result, these optimizations allow improving both performance and energy consumption radically, up to about 10.9 and 8.8 times, respectively. The impact of the Intel AVX-512 SIMD extension on the energy consumption and performance is demonstrated. Also, we discover limitations on the usability of CPU frequency scaling as a tool for balancing energy savings with admissible performance losses.

Energy-Efficient Computation Offloading in Delay-Constrained Massive MIMO Enabled Edge Network Using Data Partitioning

We study a wireless edge-computing system which allows multiple users to simultaneously offload computation-intensive tasks to multiple massive-MIMO access points, each with a collocated multi-access edge computing (MEC) server. Massive-MIMO enables simultaneous uplink transmissions from all users, significantly shortening the data offloading time compared to sequential protocols, and makes the three phases of data offloading, computing, and downloading have comparable durations. Based on this three-phase structure, we formulate a novel problem to minimize a weighted sum of the energy consumption at both the users and the MEC server under a round-trip latency constraint, using a combination of data partitioning, transmit power control and CPU frequency scaling at both the user and server ends. We design a novel nested algorithm consisting of an inner primal-dual algorithm and an outer latency-aware descent algorithm to solve this problem efficiently. Optimized solutions show that for larger requests, more data is offloaded to the MECs to reduce local computation time in order to meet the latency constraint, despite higher energy cost of wireless transmissions. Massive-MIMO channel estimation errors under pilot contamination also causes more data to be offloaded to the MECs. Compared to binary offloading, partial offloading with data partitioning is superior and leads to significant reduction in the overall energy consumption.

Energy-Aware CPU Frequency Scaling for Mobile Video Streaming

The energy consumed by video streaming includes the energy consumed for data transmission and CPU processing, which are both affected by the CPU frequency. High CPU frequency can reduce the data transmission time but it consumes more CPU energy. Low CPU frequency reduces the CPU energy but increases the data transmission time and then increases the energy consumption. In this paper, we aim to reduce the total energy of mobile video streaming by adaptively adjusting the CPU frequency. Based on real measurement results, we model the effects of CPU frequency on TCP throughput and system power. Based on these models, we propose an Energy-aware CPU Frequency Scaling (EFS) algorithm which selects the CPU frequency that can achieve a balance between saving the data transmission energy and CPU energy. Since the downloading schedule of existing video streaming apps is not optimized in terms of energy, we also propose a method to determine when and how much data to download. Through trace-driven simulations and real measurement, we demonstrate that the EFS algorithm can reduce 30 percent of energy for the Youtube app, and the combination of our download method and EFS algorithm can save 50 percent of energy than the default Youtube app.

Mildip: An energy efficient code offloading framework in mobile cloudlets

The explosive growth of connected mobile devices in Edge of Things (EoT) computing offers an opportunity to support high-quality life in smart cities but also results in an extremely high energy consumption in battery limited mobile devices. Accordingly, to increase the battery life or improve performance of mobile devices, code offloading has been proposed. Moreover, with the development of the technology about low-latency and low-power of device-to-device communication, offloading codes to surrounding mobile devices in nearby cloudlets has been widely applied. However, the existing works on the mobile cloudlets mostly focus on code distribution and scheduling but ignores the affection of the energy efficiency of CPU. According to the CPU Energy/Frequency Convexity Rule, there is an optimal clock frequency that can minimize energy consumption. In recent years, a novel processor architectures with adjustable frequency have been proposed to deliver various kinds of energy efficiency and CPU performance. Taking the processor frequency into consideration, we propose an energy-efficient code offloading framework as MilDip to offload partitioned heavy tasks to around mobile devices and to operate these tasks in the state of low CPU performance and high energy efficiency. We formulate offloading problem as a mixed-integer nonlinear optimization problem, with the goal of energy efficiency maximization. Based on the formulation, we further propose a heuristic algorithm COFS based on CPU frequency scaling. Extensive simulated experiments are provided to evaluate the performance of our method. The simulation results show that up to 77% of energy can be saved by using COFS compared to the local execution. Moreover, when enough devices surround the terminal, COFS energy consumption tends to stabilize. From the experiments, we can also conclude that COFS can save more about 20% ~ 50% energy consumption comparing to other schemes.

OKRA: optimal task and resource allocation for energy minimization in mobile edge computing systems

To cope with the computational and energy constraints of mobile devices, Mobile Edge Computing (MEC) has recently emerged as a new paradigm that provides IT and cloud-computing services at mobile network edge in close proximity to mobile devices. This paper investigates the energy consumption problem for mobile devices in a multi-user MEC system with different types of computation tasks, random task arrivals, and unpredictable channel conditions. By jointly considering computation task scheduling, CPU frequency scaling, transmit power allocation and subcarrier bandwidth assignment, we formulate it as a stochastic optimization problem aiming at minimizing the power consumption of mobile devices and to maintain the long-term stability of task queues. By leveraging the Lyapunov optimization technique, we propose an online control algorithm (OKRA) to solve the formulation. We prove that this algorithm is able to provide deterministic worst-case latency guarantee for latency-sensitive computation tasks, and balance a desirable tradeoff between power consumption and system stability by appropriately tuning the control parameter. Extensive simulations are carried out to verify the theoretical analysis, and illustrate the impacts of critical parameters to algorithm performance.

Markov Approximation for Task Offloading and Computation Scaling in Mobile Edge Computing

Mobile edge computing (MEC) provides cloud-computing services for mobile devices to offload intensive computation tasks to the physically proximal MEC servers. In this paper, we consider a multiserver system where a single mobile device asks for computation offloading to multiple nearby servers. We formulate this offloading problem as the joint optimization of computation task assignment and CPU frequency scaling, in order to minimize a tradeoff between task execution time and mobile energy consumption. The resulting optimization problem is combinatorial in essence, and the optimal solution generally can only be obtained by exhaustive search with extremely high complexity. Leveraging the Markov approximation technique, we propose a light-weight algorithm that can provably converge to a bounded near-optimal solution. The simulation results show that the proposed algorithm is able to generate near-optimal solutions and outperform other benchmark algorithms.

Performance and energy analysis of OpenMP runtime systems with dense linear algebra algorithms

In this article, we analyze performance and energy consumption of five OpenMP runtime systems over a non-uniform memory access (NUMA) platform. We also selected three CPU-level optimizations or techniques to evaluate their impact on the runtime systems: processors features Turbo Boost and C-States, and CPU Dynamic Voltage and Frequency Scaling through Linux CPUFreq governors. We present an experimental study to characterize OpenMP runtime systems on the three main kernels in dense linear algebra algorithms (Cholesky, LU, and QR) in terms of performance and energy consumption. Our experimental results suggest that OpenMP runtime systems can be considered as a new energy leverage, and Turbo Boost, as well as C-States, impacted significantly performance and energy. CPUFreq governors had more impact with Turbo Boost disabled, since both optimizations reduced performance due to CPU thermal limits. An LU factorization with concurrent-write extension from libKOMP achieved up to 63% of performance gain and 29% of energy decrease over original PLASMA algorithm using GNU C compiler (GCC) libGOMP runtime.

Dynamic Frequency Scaling Regarding Memory for Energy Efficiency of Embedded Systems

Memory significantly affects the power consumption of embedded systems as well as performance. CPU frequency scaling for power management could fail in optimizing the energy efficiency without considering the memory access. In this paper, we analyze the power consumption and energy efficiency of an embedded system that supports dynamic scaling of frequency for both CPU and memory access. The power consumption of the CPU and the memory is modeled to show that the memory access rate affects the energy efficiency and the CPU frequency selection. Based on the power model, a method for frequency selection is presented to optimize the power efficiency which is measured using Energy-Delay Product (EDP). The proposed method is implemented and tested on a commercial smartphone to achieve about 3.3% - 7.6% enhancement comparing with the power management policy provided by the manufacturer in terms of EDP.

Accelerating next generation sequencing data analysis with system level optimizations

Next generation sequencing (NGS) data analysis is highly compute intensive. In-memory computing, vectorization, bulk data transfer, CPU frequency scaling are some of the hardware features in the modern computing architectures. To get the best execution time and utilize these hardware features, it is necessary to tune the system level parameters before running the application. We studied the GATK-HaplotypeCaller which is part of common NGS workflows, that consume more than 43% of the total execution time. Multiple GATK 3.x versions were benchmarked and the execution time of HaplotypeCaller was optimized by various system level parameters which included: (i) tuning the parallel garbage collection and kernel shared memory to simulate in-memory computing, (ii) architecture-specific tuning in the PairHMM library for vectorization, (iii) including Java 1.8 features through GATK source code compilation and building a runtime environment for parallel sorting and bulk data transfer (iv) the default ’on-demand’ mode of CPU frequency is over-clocked by using ’performance-mode’ to accelerate the Java multi-threads. As a result, the HaplotypeCaller execution time was reduced by 82.66% in GATK 3.3 and 42.61% in GATK 3.7. Overall, the execution time of NGS pipeline was reduced to 70.60% and 34.14% for GATK 3.3 and GATK 3.7 respectively.

Performance-Based Pricing in Multi-Core Geo-Distributed Cloud Computing

New pricing policies are emerging where cloud providers charge resource provisioning based on the allocated CPU frequencies. As a result, resources are offered to users as combinations of different performance levels and prices which can be configured at runtime. With such new pricing schemes and the increasing energy costs in data centres, balancing energy savings with performance and revenue losses is a challenging problem for cloud providers. CPU frequency scaling can be used to reduce power dissipation, but also impacts VM performance and therefore revenue. In this paper, we firstly propose a non-linear power model that estimates power dissipation of a multi-core PM and secondly a pricing model that adjusts the pricing based on the VM's CPU-boundedness characteristics. Finally, we present a cloud controller that uses these models to allocate VMs and scale CPU frequencies of the PMs to achieve energy cost savings that exceed service revenue losses. We evaluate the proposed approach using simulations with realistic VM workloads, electricity price and temperature traces and estimate energy savings of up to 14.57%.

Multimedia Content Delivery with Network Function Virtualization: The Energy Perspective

Multimedia now accounts for the largest share of all Internet traffic, highlighted by its volume, variety, multicast nature, and QoS constraints. Downstream toward consumers, multimedia traffic can traverse through middleboxes, undergoing additional data processing imposed by content providers and distributors. With the advent of network function virtualization (NFV), middleboxes are progressively embedded in off-the-shelf, general-purpose servers. Despite the benefits, NFV can incur an undue amount of energy consumption during high packet forwarding. In this article, the authors investigate how switching to state-of-the-art NFV products for multimedia content delivery can result in significant energy costs. They identify energy inefficiency in the NFV data plane, which can be exacerbated if not handled properly. They outline a power management framework that considers characteristics of multimedia traffic and exploits CPU frequency scaling to save energy. This article is part of a special issue on advancing multimedia distribution.

An Energy Efficient Code Offloading Approach for Mobile Cloud Computing

Objectives: This research work proposes a methodology for estimating the energy consumption of tasks by considering processor and memory usage. Methods: To facilitate energy efficiency in CPU, the mobile devices may be operated at different frequencies during the execution of tasks. This research work applies CPU frequency scaling as its base to achieve energy efficiency. Besides, it also considers energy consumption during memory access while making the offloading decision. Findings: The proposed approach uses energy consumption during computation as well as memory access as its metric to conceive the offloading decision. Additionally, the proposed energy model is simulated and the results are concluded that there is a considerable amount of energy saving in mobile devices due to computation offloading to nearby mobile devices or cloud resources. Applications: To save energy decide which application in mobile as energy consume considered as model. The models work as code offloading in MATLAB and determine by two level genetic algorithms. The efficiency of proposed model is evaluated by a simulation and average energy result can be concluded for a mobile device.

Performance Analysis of MTD64, our Tiny Multi-Threaded DNS64 Server Implementation: Proof of Concept

In the current stage of IPv6 deployment, the combination of DNS64 and NAT64 is an important IPv6 transition technology, which can be used to enable IPv6 only clients to communicate with IPv4 only servers. In addition to the existing free software DNS64 implementations, we proposed a tiny multithreaded one, MTD64.In this paper, the performance of MTD64 is measured and compared to that of the industry standard BIND in order to check the correctness of the design concepts of MTD64, especially of the one that we use a new thread for each request. For the performance measurements, our earlier proposed dns64perf program is enhanced as dns64perf2, which one is also documented in this paper. We found that MTD64 seriously outperformed BIND and hence our design principles may be useful for the design of a high performance production class DNS64 server. As an additional test, we have also examined the effect of dynamic CPU frequency scaling to the performance of the implementations.

Design and implementation of energy-aware application-specific CPU frequency governors for the heterogeneous distributed computing systems

This paper deals with the design of application-specific energy-aware CPU frequency scaling mechanisms. The proposed customized CPU controllers may optimize performance of data centers in which diverse tasks are allocated to servers with different characteristics. First, it is demonstrated that server power usage can be accurately estimated based on the measurements of CPU power consumption read from the model specific registers (MSRs). Next, a benchmarking methodology derived from the RFC2544 specification is proposed that allows to identify models of CPU workload dynamics. Finally, it is demonstrated how the identified models can be applied in the design of customized energy-aware controllers that dynamically adjust CPU frequency to the application-specific workload patterns. According to the results of experimental studies the customized controllers may outperform standard general-purpose governors of the Linux kernel both in terms of reachable server performance and power saving capabilities.

DREAM: Dynamic Resource and Task Allocation for Energy Minimization in Mobile Cloud Systems

To cope with increasing energy consumption in mobile devices, the mobile cloud offloading has received considerable attention from its ability to offload processing tasks of mobile devices to cloud servers, and previous studies have focused on single type tasks in fixed network environments. However, real network environments are spatio-temporally varying, and typical mobile devices have not only various types of tasks, e.g., network traffic, cloud offloadable/nonoffloadable workloads but also capabilities of CPU frequency scaling and network interface selection between WiFi and cellular. In this paper, we first jointly consider the following three dynamic problems in real mobile environments: 1) cloud offloading policy, i.e., determining to use local CPU resources or cloud resources; 2) allocation of tasks to transmit through networks and to process in local CPU; and 3) CPU clock speed and network interface controls. We propose a DREAM algorithm by invoking the Lyapunov optimization and mathematically prove that it minimizes CPU and network energy for given delay constraints. Trace-driven simulation based on real measurements demonstrates that DREAM can save over 35% of total energy than existing algorithms with the same delay. We also design DREAM architecture and demonstrate the applicability of DREAM in practice.

A reconfigurable framework for compositional schedulability and power analysis of hierarchical scheduling systems with frequency scaling

This paper presents a compositional framework for the modeling and analysis of hierarchical scheduling systems. We consider both schedulability and energy consumption of individual components, while analyzing a single core setting with a voltage frequency scaling CPU. According to the CPU frequency scaling, each task has a set of different execution times. Thus, the energy consumption of the whole system varies from one execution to another.We analyze each component individually by checking the feasibility of its workload against both the CPU availability and energy consumption constraints of such a component. Our periodic task model considers both static and dynamic priorities together with preemptive and non-preemptive behaviors. The models are realized using different forms of Hybrid Automata, all of which are analyzed using variants of Uppaal. The CPU frequencies, task behavior and scheduling policies used in each component are some of the reconfigurable parameters of the system. Finally, we demonstrate the applicability and scalability of our framework by analyzing the schedulability and power consumption of an avionics system.

온디맨드 거버너 정책에 따른 작업 스케줄링 기법의 성능 평가

The ondemand governor used in android-based smartphone platforms is a CPU frequency scaling technique. The ondemand governor sets the CPU operating frequency depending on the CPU utilization rate. Job scheduling affects the CPU utilization rate. The power consumption is proportional to the value of operating frequency. Consequently, CPU frequency scaling and CPU utilization rate have an effect on power consumption in a smartphone. In this paper, we evaluated the performance of job scheduling techniques incorporating the ondemand governor in terms of CPU utilization, power consumption, and job deadline miss ratio.

Energy‐efficient CPU frequency control for the Linux system

SummaryEfficiency of energy usage in computing systems improves, however, still not at the rate matching the climbing demand for computing capacity. To address this urging problem, computing elements of the latest generation, that is, CPUs/graphics processing units, memory units, and network interface cards, have been designed to operate in multiple modes with differentiated energy consumption levels. Mode switching and high‐frequency performance monitoring functions have also been exposed by co‐designed abstract programming interfaces. The challenge of energy‐efficient computing is to develop hardware control mechanisms taking advantage of the new capabilities. This paper aims at giving an insight into the structure of optimal energy‐aware CPU frequency scaling rules. It gives a characterization of solutions to the optimal control problem of energy‐efficient real‐time packet inspection performed by a Linux server. A class of CPU frequency switching rules, exploiting dynamic voltage and frequency scaling mechanisms, is constructed based on experimentally identified model of server operations. The control rules are demonstrated to outperform the default CPU frequency scaling governor for the Linux kernel, both in terms of achievable power savings and service quality. Copyright © 2015 John Wiley & Sons, Ltd.

Governing energy consumption in Hadoop through CPU frequency scaling: An analysis

With increasingly inexpensive storage and growing processing power, the cloud has rapidly become the environment of choice to store and analyze data for a variety of applications. Most large-scale data computations in the cloud heavily rely on the MapReduce paradigm and on its Hadoop implementation. Nevertheless, this exponential growth in popularity has significantly impacted power consumption in cloud infrastructures. In this paper, we focus on MapReduce processing and we investigate the impact of dynamically scaling the frequency of compute nodes on the performance and energy consumption of a Hadoop cluster. To this end, a series of experiments are conducted to explore the implications of Dynamic Voltage and Frequency Scaling (DVFS) settings on power consumption in Hadoop clusters. By enabling various existing DVFS governors (i.e., performance, powersave, ondemand, conservative and userspace) in a Hadoop cluster, we observe significant variation in performance and power consumption across different applications: the different DVFS settings are only sub-optimal for several representative MapReduce applications. Furthermore, our results reveal that the current CPU governors do not exactly reflect their design goal and may even become ineffective to manage the power consumption in Hadoop clusters. This study aims at providing a clearer understanding of the interplay between performance and power management in Hadoop clusters and therefore offers useful insight into designing power-aware techniques for Hadoop systems.