Existing Power Models Research Articles

McPAT-Calib: A RISC-V BOOM Microarchitecture Power Modeling Framework

Power efficiency has become a nonneglected issue of modern CPUs. Therefore, accurate and robust power models are highly demanded in academia and industry. However, it is hard for existing power models to balance modeling speed, generality, and accuracy well. This article introduces McPAT-Calib, a microarchitecture power modeling framework, which combines McPAT with machine learning (ML) calibration and active learning (AL) sampling. McPAT-Calib can quickly and accurately estimate the power of different benchmarks executed on different CPU configurations, and provide an effective evaluation tool for the early design stage. First, McPAT-7nm is introduced to support the preliminary analytical power modeling for the 7-nm technology node. Then, a wide range of modeling features are identified, and automatic feature selection and advanced nonlinear regression are used to calibrate the McPAT-7nm modeling results, greatly improving the accuracy. Moreover, a novel AL approach termed power greedy sampling (PowerGS) embedded with domain knowledge is leveraged to reduce the modeling cost effectively. We use up to 15 configurations of the RISC-V Berkeley out-of-order machine (BOOM) along with 80 benchmarks, targeting 7-nm technology, to extensively evaluate McPAT-Calib. Compared with state-of-the-art (SOTA) microarchitecture power models, McPAT-Calib can reduce the mean absolute percentage error (MAPE) under different cross-validation (CV) strategies by 3.64%–6.14% (absolute reduction). Meanwhile, PowerGS is superior to the existing AL approaches, which can significantly reduce the demand for labeled samples to speed up model construction. The effectiveness of the overall modeling and estimation flow with AL sampling has also been verified.

Phase-Based Accurate Power Modeling for Mobile Application Processors

Modern mobile application processors are required to execute heavier workloads while the battery capacity is rarely increased. This trend leads to the need for a power model that can analyze the power consumed by CPU and GPU at run-time, which are the key components of the application processor in terms of power savings. We propose novel CPU and GPU power models based on the phases using performance monitoring counters for smartphones. Our phase-based power models employ combined per-phase power modeling methods to achieve more accurate power consumption estimations, unlike existing power models. The proposed CPU power model shows estimation errors of 2.51% for ARM Cortex A-53 and 1.97% for Samsung M1 on average, and the proposed GPU power model shows an average error of 8.92% for the Mali-T880. In addition, we integrate proposed CPU and GPU models with the latest display power model into a holistic power model. Our holistic power model can estimate the smartphone′s total power consumption with an error of 6.36% on average while running nine 3D game benchmarks, improving the error rate by about 56% compared with the latest prior model.

An adaptive workload-aware power consumption measuring method for servers in cloud data centers

As cloud computing technologies and applications develop rapidly in recent years, the quantity and size of cloud datacenters have been ever-increasing, making the overconsumption of energy in datacenters become a widespread concern. To reduce the energy cost by servers, we must first build an accurate power model to achieve flexible, device-free power consumption measuring. However, most of the previous work related to server power modeling solely apply to the server and virtual machine levels, and the existing power models fail to take into account the heterogeneity in workload. Therefore, we first propose separate power consumption models based on the distinction of workload types including CPU-intensive, I/O-intensive, memory-intensive, and mixed workload. Then, we present an adaptive workload-aware power consumption measuring method (WSPM) for cloud servers. Our method proactively selects an appropriate power model for the upcoming workload through workload clustering, forecasting and classification, which are implemented using K-means, ARIMA, and threshold-based methods, respectively. We conducted several experiments to evaluate the performance of the key components of our method. The result shows: (1) the accuracy of our future workload forecasting on real traces of requests to our servers, (2) the accuracy of the power consumption measured by WSPM, and (3) the effectiveness of our workload-aware method in reducing real-time power estimation lag. Overall, the proposed method simplifies power modeling under diverse workloads without losing accuracy, making it a general and highly available solution for cloud data centers.

Mobile AMOLED Display Power Model Considering I–R Drop in Smartphones

In order to efficiently manage display power, there is a need for a system that can accurately predict the power consumed according to the displayed image. Existing AMOLED display power models do not consider the effects of I-R drop, which is a voltage drop along the VDD line. Therefore, when the existing power model is applied to a recent AMOLED display panel, which has a large physical size and uses a large current, power consumption cannot be accurately predicted. In this article, we propose a new model to compensate for the I-R drop and present a more accurate method for predicting AMOLED display power consumption by applying it to existing AMOLED display power models. To verify this, we conducted extensive experiments for various image databases on AMOLED display panels of Samsung Galaxy S7 and S9. When comparing the proposed display power model connected with the I-R drop model with conventional power models, it is confirmed that the proposed power model improves the accuracy by up to 70.3% over them.

Modeling and Assessing the Power Consumption Behavior of Sensor Nodes using Petri Nets

Power Consumption is an influential and important concern in Wireless Sensor Networks (WSNs). Assessing and determining the impact of the power behavior of sensor nodes is an inclusive concern that should be executed at the network pre-deployment phase rather than post-deployment phase. Providing an accurate power modeling improves the development process of WSN applications and protocols. This paper introduces the use of Colored Petri Nets (CPNs) to model the power behavior and relations among different sensor node components when operated in an event driven environment. The objective is to figure out the overall power behavior of the nodes by considering power consumed accompanied with different states transition at a specific packets arrival and service rate values. Colored Petri Nets is a modeling language employed for validating and evaluating concurrent and distributed systems. The introduced model is beneficial since it provides the network designer with a way to determine surrogate designs or to check the compatibility of an existing power model behavior. The proposed model has been validated through the comparison of the power behaviors of two sensor nodes Mica2 and Telos. The results demonstrate the efficiency of using the proposed model to draw and analyze the WSNs energy consumption behavior.

A mechanics based prediction model for tool wear and power consumption in drilling operations and its applications

We present a mechanics based model for predicting the power consumption of drilling operations. Different from existing power models in machining that ignore the tool wear, our model takes into full consideration the tool wear which is particularly pronounced in drilling and causes extra power consumption. For any given spindle speed n and feed rate f, our model establishes the relationship between the length of drill and the total power consumption as well as the amount of tool wear. With this prediction model established, we can then optimize the drilling parameters (n, f) towards different objectives, such as the two applications reported in this paper – to minimize the average power consumption per unit length of drill and to maximize the tool usage before its replacement. Physical drilling experiments of the proposed power prediction model and its two optimization applications are also reported in this paper which have validated the accuracy of the model and convincingly demonstrated its efficacy in deciding optimal drilling parameters (n, f) for energy minimization and other objectives.

Accurate power modeling of modern mobile application processors

The power modeling of mobile application processors (APs) is a challenging task due to their complexity. The existing power models and their associated devices have mostly been made obsolete by rec...

Accurate power modeling of modern mobile application processors

The power modeling of mobile application processors (APs) is a challenging task due to their complexity. The existing power models and their associated devices have mostly been made obsolete by recent hardware developments. In this paper, we propose an enhanced power model used in modern mobile devices. The model accurately estimates the power consumption of AP component and utilizes the runtime usage information of each hardware component. We evaluated the model accuracy using various benchmarks, as well as popular smartphone applications with multiple devices that employ different APs. The evaluation shows that our model achieves the mean absolute percentage error (MAPE) of 5.1%.

Impact of Energy-Efficient and Eco-Friendly Green Computing

the abrupt advancement in technology, every business organization aspires to migrate to cloud. The technology of cloud computing makes use of the huge data centers that present multiple issues such as extensive amounts of energy consumption, dissipation of lots of heat and methane, carbon dioxide, etc. like deadly green-house compounds and gases. In order to solve all the above stated issues of cloud computing, green computing came out as an improvement over the conventional cloud technology with energy-efficiency, security and environmental-friendliness. The two of the Linear power model and the Low Powerblade model are the existing ones, which employ four Virtual Machine scheduling algorithms for the calculation of consumption of power, are not very eco-friendly and energy efficient. These limitations of the two power models call for the requirement of a better and more efficient power model. Therefore, the research work is motivated to devise such a power model that ensures energy-efficient results and proves it environment-friendly when applied on the data centers in a secure way when compared with the two existing power models on the GreenCloud Simulator.

Accurate power model for mobile AMOLED displays

The display is a dominant power-consuming component in mobile devices. The active matrix organic light-emitting diode (AMOLED) display has been popular in mobile devices and its accurate power model is crucial for efficient power management. Existing power models do not consider the behaviours of the driving hardware components of AMOLED display modules. A new power model is proposed which considers the efficiency of an AMOLED DC–DC converter and the dependency of the driving circuits in a commercial AMOLED display module, and thus estimates an accurate power consumption value. Through extensive experiments using three well-known image databases, the proposed power model is found to have lower error rates than the existing pixel-based one for all the databases. Especially, for bright images the error rate of the new model amounts to only 0.59%, whereas that of the existing one is 8.48%.

Development of a variable speed compressor power model for single-stage packaged DX rooftop units

Variable speed technology was applied to compressors for rooftop units (RTUs) and heat pumps in recent years. Compressor power is a critical parameter to evaluate system performance and conduct fault detection and diagnosis (FDD). However, existing power models often use numerous parameters related to the compressor geometrical dimensions and refrigerant physical properties which makes the existing models hard to implement in the field. In this paper, a semi-theoretical compressor power model was developed for single-stage RTUs equipped with a variable speed compressor and a variable speed indoor fan, based on the theoretical analysis and experimental studies. Under the normal conditions, the compressor power is correlated to the outside air temperature and compressor speed with a relative error of ±8%. This model can be employed in the field to develop a compressor power baseline for real-time FDD on the Direct Expansion (DX) RTUs.

무선 통신 시스템의 전력 모델을 이용한 비트당 최소 에너지

본 논문은 RF power model과 주파수 대역의 특성을 이용한 비트 당 에너지와 전송량과의 관계를 시스템 대역폭의 변화에 따라 분석한 논문이다. 기존에 제안된 RF power model은 각각의 디바이스의 소모 전력을 수식적으로 표현한 것이다. 이 전력 모델에 고려된 요소는 시스템의 전송 대역과 PAR, 데이터 전송량, 변조 레벨, 전송, 전송 거리 등이다. 본 논문에서는 이러한 영향을 고려하여 RF power model과 주파수 대역의 특성을 이용한 비트당 에너지와 전송량의 관계를 시스템 대역폭의 변화에 따라 분석하였다. Shannon capacity 공식과 신호의 SNR에 대한 식, 그리고 RF power model의 소모 전력을 이용하여 해당 주파수에서의 소모 전력을 구하고, Gbps급 데이터 속도에 따른 비트 당 에너지의 최소 값을 찾기 위한 시뮬레이션을 진행하였다. In this paper, we analyze the relationship between energy per bit and the data rate with the variation of the system bandwidth. A existing power model is mathematical model to express power consumption of each device. In this paper, we have to investigate the system level energy model for the RF front-end of a wireless transceiver. Also, the effects of the signal bandwidth, PAR, date rate, modulation level, transmission distance, specific attenuation of frequency band, and the signal center frequency on the RF front-end energy consumption and system capacity are considered. Eventually, we analyze the relationship between energy per bit and the data rate with the variation of the system bandwidth so that we simulate the minimum energy per bit in the several Gbps data rate using Shannon capacity theory.

SPAN: A software power analyzer for multicore computer systems

Abstract Understanding the power dissipation behavior of an application/workload is the key to writing power-efficient software and designing energy-efficient computer systems. Power modeling based on performance monitoring counters (PMCs) is an effective approach to analyze and quantify power dissipation behaviors on a real computer system. One of the potential benefits is that software developers are able to optimize the power behavior of an application by adjusting its source code implementations. However, it is challenging to relate power dissipation to the execution of specific segments of source code directly. In addition, existing power models need to be further investigated by reconsidering multicore architecture processors with on-chip shared resources. Therefore, we need to adjust PMC-based power models from the developers’ perspective, and reevaluate them on multicore computer systems. In this paper, followed by a detailed classification of previous efforts on power profiling, we propose a two-level power model that estimates per-core power dissipation on chip multiprocessor (CMP) on-the-fly by using only one PMC and frequency information from CPUs. The model attempts to satisfy the basic requirements from developer point of view: simplicity and applicability. Based on this model, we design and implement SPAN, a software power analyzer, to identify power behavior associated with source code. Given an application, SPAN is able to determine its power dissipation rate at the function-block level. We evaluate both the power model and SPAN on two general purpose multicore computer systems. The experimental results based on SPEC2008Cjvm benchmark suite show the average error rate of 5.40% across one core to six core validation. We also verify SPAN using the FT benchmark from NAS parallel benchmark suite and a synthetic workload. The overall estimated error of SPAN is under 3.00%.