Measuring energy consumption using EML (energy measurement library)

Cabrera, Alberto; Almeida, Francisco; Arteaga, Javier Cabrera; Blanco, Vicente

doi:10.1007/s00450-014-0269-5

Cited by 25 publications

(9 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Usually this period must be at least a couple of seconds long. Linearity: the indicator value is directly proportional to performance. Reliability: the value of the indicator is computed using measurable facts and there is no subjective component included in the measurement. Efficiency: computing the indicator is lightweight and requires little computing power. This is due to the fact that the utilization of individual components can directly be read from counters of the operation system (OS) while the energy consumption of these components can be estimated using OS counters (Economou et al , 2006; Povoa et al , 2013) or collected, for example, through the energy measurement library (Cabrera et al , 2015). However, the necessary extraction and/or estimation of the different variables (i.e.…”

Section: Discussionmentioning

confidence: 99%

“…Efficiency: computing the indicator is lightweight and requires little computing power. This is due to the fact that the utilization of individual components can directly be read from counters of the operation system (OS) while the energy consumption of these components can be estimated using OS counters (Economou et al , 2006; Povoa et al , 2013) or collected, for example, through the energy measurement library (Cabrera et al , 2015). However, the necessary extraction and/or estimation of the different variables (i.e.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

EEUI: a new measure to monitor and manage energy efficiency in data centers

Abaunza

Hameri

Niemi

2018

IJPPM

View full text Add to dashboard Cite

Purpose Data centers (DCs) are similar to traditional factories in many aspects like response time constraints, limited capacity, and utilization levels. Several indicators have been developed to monitor and compare productivity in manufacturing. However, in DCs most used indicators focus on technical aspects of infrastructure, not efficiency of operations. The purpose of this paper is to rely on operations management to define a commensurate and proportionate DC performance indicator: the energy-efficient utilization indicator (EEUI). EEUI makes objective and comparative assessment of efficiency possible independently of the operating environment and its constraints. Design/methodology/approach The authors followed a design science approach, which follows the practitioner’s initial steps for finding solutions to business relevant problems prior to theory building. Therefore, this approach fits well with this research, as it is primarily motivated by business and management needs. EEUI combines both the amount of energy consumed by different components and their current energy efficiency (EE). It reaches its highest value when all server components are optimally loaded in EE sense. The authors tested EEUI by collecting data from three scientific DCs and performing controlled laboratory tests. Findings The results indicate that the optimization of EEUI makes it possible to run computing resources more efficiently. This leads to a higher EE and throughput of the DC while reducing the carbon footprint associated to DC operations. Both energy-related costs and the total cost of ownership are consequently reduced, since the amount of both energy and hardware resources needed decrease, while improving DC sustainability. Practical implications In comparison with current DC operations, the results imply that using the EEUI could help increase the EE of DCs. In order to optimize the proposed EEUIs, DC managers and operators should use resource management policies that increase the resource usage variation of the jobs being processed in the same computing resources (e.g. servers). Originality/value The paper provides a novel approach to monitor the EE at which computing resources are used. The proposed indicator not only considers the utilization levels at which server components are used but also takes into account their EE and energy proportionality.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

EEUI: a new measure to monitor and manage energy efficiency in data centers

Abaunza

Hameri

Niemi

2018

IJPPM

View full text Add to dashboard Cite

show abstract

“…In order to illustrate the proposed methodology, this section presents an in- Energy measurements were gathered using the EML [28], a driver based energy measurement library. In our experimentation, we used the appropriate drivers to extract total energy metrics from the RAPL interface.…”

Section: Pareto Front Based Analysismentioning

confidence: 99%

Energy-efficient power cap configurations through Pareto Front Analysis and Machine Learning Categorization

Cabrera

Almeida

Castellanos–Nieves

et al. 2023

Preprint

View full text Add to dashboard Cite

The growing demand for more computational resources has increased the overall energy consumption in computer systems. To support the increasing requirements, power and energy consumption have to be considered as a constraint to execute software. Modern architectures provide tools to manage directly the power constraints of a system. The Intel Power Cap is a relatively recent tool developed to offer fine control of power usage to users at a central processing unit (CPU) level. We propose a methodology to analyze the performance and the energy efficiency trade-offs using this power cap technology for a given algorithm. We extract a Pareto front for the multi-objective performance and energy problem to represent multiple feasible configurations for both objectives. We perform an extensive experimentation to categorize the different algorithms to reduce the total amount of optimal power cap configurations. We propose the use of Machine Learning (ML) clustering techniques to categorize any algorithm in the target architecture.The use of Machine Learning allows to automate the procedure and simplifies the effort required to solve the optimization problem.We present a practical case where we categorize the kernels using ML techniques, with the option to include a new algorithm into an already existing categorization.

show abstract

“…They find that the greenest computers are often small-scale deployments and they propose a new energy-efficiency metric that also takes into account system scale. Cabrera et al [30] present the Energy Measurement Library (EML): a framework that standardizes energy measurements across different devices (such as Intel CPUs, NVIDIA GPUs, and Power Distribution Units (PDU). Shoukourian et al [164] propose the PowerDAM toolset for collecting sensor data in HPC clusters and correlating sensor data with resource management data.…”

Section: Modeling and Toolsmentioning

confidence: 99%

The Landscape of Exascale Research

et al. 2020

View full text Add to dashboard Cite

The next generation of supercomputers will break the exascale barrier. Soon we will have systems capable of at least one quintillion (billion billion) floating-point operations per second (10 18 FLOPS). Tremendous amounts of work have been invested into identifying and overcoming the challenges of the exascale era. In this work, we present an overview of these efforts and provide insight into the important trends, developments, and exciting research opportunities in exascale computing. We use a three-stage approach in which we (1) discuss various exascale landmark studies, (2) use data-driven techniques to analyze the large collection of related literature, and (3) discuss eight research areas in depth based on influential articles. Overall, we observe that great advancements have been made in tackling the two primary exascale challenges: energy efficiency and fault tolerance. However, as we look forward, we still foresee two major concerns: the lack of suitable programming tools and the growing gap between processor performance and data bandwidth (i.e., memory, storage, networks). Although we will certainly reach exascale soon, without additional research, these issues could potentially limit the applicability of exascale computing.

show abstract

Measuring energy consumption using EML (energy measurement library)

Cited by 25 publications

References 20 publications

EEUI: a new measure to monitor and manage energy efficiency in data centers

EEUI: a new measure to monitor and manage energy efficiency in data centers

Energy-efficient power cap configurations through Pareto Front Analysis and Machine Learning Categorization

The Landscape of Exascale Research

Contact Info

Product

Resources

About