Does lean imply green?: a study of the power performance implications of Java runtime bloat

The presence of software bloat in large flexible software systems can hurt energy efficiency. However, identifying and mitigating bloat is fairly effort intensive. To enable such efforts to be directed where there is a substantial potential for energy savings, we investigate the impact of bloat on power consumption under different situations. We conduct the first systematic experimental study of the joint power-performance implications of bloat across a range of hardware and software configurations on modern server platforms. The study employs controlled experiments to expose different effects of a common type of Java runtime bloat, excess temporary objects, in the context of the SPECPower_ssj2008 workload. We introduce the notion of equi-performance power reduction to characterize the impact, in addition to peak power comparisons. The results show a wide variation in energy savings from bloat reduction across these configurations. Energy efficiency benefits at peak performance tend to be most pronounced when bloat affects a performance bottleneck and non-bloated resources have low energy-proportionality. Equi-performance power savings are highest when bloated resources have a high degree of energy proportionality. We develop an analytical model that establishes a general relation between resource pressure caused by bloat and its energy efficiency impact under different conditions of resource bottlenecks and energy proportionality. Applying the model to different "what-if" scenarios, we predict the impact of bloat reduction and corroborate these predictions with empirical observations. Our work shows that the prevalent software-only view of bloat is inadequate for assessing its power-performance impact and instead provides a full systems approach for reasoning about its implications.

Prediction-Based Power-Performance Adaptation of Multithreaded Scientific Codes

Computing has recently reached an inflection point with the introduction of multicore processors. On-chip thread-level parallelism is doubling approximately every other year. Concurrency lends itself naturally to allowing a program to trade performance for power savings by regulating the number of active cores; however, in several domains, users are unwilling to sacrifice performance to save power. We present a prediction model for identifying energy-efficient operating points of concurrency in well-tuned multithreaded scientific applications and a runtime system that uses live program analysis to optimize applications dynamically. We describe a dynamic phase-aware performance prediction model that combines multivariate regression techniques with runtime analysis of data collected from hardware event counters to locate optimal operating points of concurrency. Using our model, we develop a prediction-driven phase-aware runtime optimization scheme that throttles concurrency so that power consumption can be reduced and performance can be set at the knee of the scalability curve of each program phase. The use of prediction reduces the overhead of searching the optimization space while achieving near-optimal performance and power savings. A thorough evaluation of our approach shows a reduction in power consumption of 10.8 percent, simultaneous with an improvement in performance of 17.9 percent, resulting in energy savings of 26.7 percent.

Improving the ultra-low-power performance of IEEE 802.15.6 by adaptive synchronisation

In ultra-low data rate wireless sensor networks (WSNs) waking up just to listen to a beacon every superframe can be a major waste of energy. This study introduces MedMAC, a medium access protocol for ultra-low data rate WSNs that achieves significant energy efficiency through a novel synchronisation mechanism. The new draft IEEE 802.15.6 standard for body area networks includes a sub-class of applications such as medical implantable devices and long-term micro miniature sensors with ultra-low power requirements. It will be desirable for these devices to have 10 years or more of operation between battery changes, or to have average current requirements matched to energy harvesting technology. Simulation results are presented to show that the MedMAC allows nodes to maintain synchronisation to the network while sleeping through many beacons with a significant increase in energy efficiency during periods of particularly low data transfer. Results from a comparative analysis of MedMAC and IEEE 802.15.6 MAC show that MedMAC has superior efficiency with energy savings of between 25 and 87 for the presented scenarios. © 2011 The Institution of Engineering and Technology.

Optimization of FRT Active Power Performance of a DFIG during Transient Grid Faults

Optimal fault ride-through (FRT) conditions for a doubly-fed induction generator (DFIG) during a transient grid fault are analyzed with special emphasis on improving the active power generation profile. The transition states due to crowbar activation during transient faults are investigated to exploit the maximum power during the fault and post-fault period. It has been identified that operating slip, severity of fault and crowbar resistance have a direct impact on the power capability of a DFIG, and crowbar resistance can be chosen to optimize the power capability. It has been further shown that an extended crowbar period can deliver enhanced inertial response following the transient fault. The converter protection and drive train dynamics have also been analyzed while choosing the optimum crowbar resistance and delivering enhanced inertial support for an extended crowbar period.

Power performance with gated clocks of a pipelined Cordic Core

This paper presents the evaluation in power consumption of gated clocks pipelined circuits with different register configurations in Virtex-based FPGA devices. Power impact of a gated clock circuitry aimed at reducing flip-flops output rate at the bit level is studied. Power performance is also given for pipeline stages based on the implementation of a double edge-triggered flip-flop. Using a pipelined Cordic Core circuit as an example, this study did not find evidence in power benefits either when gated clock at the bit-level or double-edge triggered flip-flops used when synthesized with FPGA logic resources.

Power performance evaluation of an electric home fan with triac-based automatic speed control system

In order to provide a low cost system of thermal comfort, a common model of home fan, 40 cm diameter size, had its manual four-button control system replaced by an automatic speed control. The new control system has a temperature sensor feeding a microcontroller that, by using an optic coupling, DIAC or TRIAC-based circuit, varies the RMS value of the fan motor input voltage and its speed, according to the room temperature. Over a wide range of velocity, the fan net power and the motor fan input power were measured working under both control system. The temperature of the motor stator and the voltage waveforms were observed too. Measured values analysis showed that the TRIAC-based control system makes the fan motor work at a very low power factor and efficiency values. The worst case is at low velocity range where the higher fan motor stator temperatures were registered. The poor power factor and efficiency and the harmonics signals inserted in the motor input voltage wave by the TRIAC commutation procedure are correlated.

Postactivation potentiation: effect of various recovery intervals on bench press power performance

Ferreira, SLA, Panissa, VLG, Miarka, B, and Franchini, E. Postactivation potentiation: effect of various recovery intervals on bench press power performance. J Strength Cond Res 26(3): 739-744, 2012-Postactivation potentiation (PAP) is a strategy used to improve performance in power activities. The aim of this study was to determine if power during bench press exercise was increased when preceded by 1 repetition maximum (1RM) in the same exercise and to determine which time interval could optimize PAP response. For this, 11 healthy male subjects (age, 25 +/- 4 years; height, 178 +/- 6 cm; body mass, 74 +/- 8 kg; bench press 1RM, 76 +/- 19 kg) underwent 6 sessions. Two control sessions were conducted to determine both bench press 1RM and power (6 repetitions at 50% 1RM). The 4 experimental sessions were composed of a 1RM exercise followed by power sets with different recovery intervals (1, 3, 5, and 7 minutes), performed on different days, and determined randomly. Power values were measured via Peak Power equipment (Cefise, Nova Odessa, Sao Paulo, Brazil). The conditions were compared using an analysis of variance with repeated measures, followed by a Tukey test. The significance level was set at p < 0.05. There was a significant increase in PAP in concentric contractions after 7 minutes of recovery compared with the control and 1-minute recovery conditions (p < 0.05). Our results indicated that 7 minutes of recovery has generated an increase in PAP in bench press and that such a strategy could be applied as an interesting alternative to enhance the performance in tasks aimed at increasing upper-body power performance.

Feasibility of combined wind-wave energy platforms

The European Union has set out an ambitious 20% target for renewable energy use by 2020. It is expected that this will be met mainly by wind energy. Looking towards 2050, reductions in greenhouse gas emissions of 80-95% are to be sought. Given the issues securing this target in the transport and agriculture sectors, it may only be possible to achieve this target if the power sector is carbon neutral well in advance of 2050. This has permitted the vast expansion of offshore renewables, wind, wave and tidal energy. Offshore wind has undergone rapid development in recent years however faces significant challenges up to 2020 to ensure commercial viability without the need for government subsidies. Wave energy is still in the very early stages of development so as yet there has been no commercial roll out. As both of these technologies are to face similar challenges in ensuring they are a viable alternative power generation method to fossil fuels, capitalising on the synergies is potentially a significant cost saving initiative. The advent of hybrid solutions in a variety of configurations is the subject of this thesis. A singular wind-wave energy platform embodies all the attributes of a hybrid system, including sharing space, transmission infrastructure, O&M activities and a platform/foundation. This configuration is the subject of this thesis, and it is found that an OWC Array platform with multi-MegaWatt wind turbines is a technically feasible, and potentially an economically feasible solution in the long term. Methods of design and analysis adopted in this thesis include numerical and physical modelling of power performance, structural analysis, fabrication cost modelling, simplified project economic modelling and time domain reliability modelling of a 210MW hybrid farm. The application of these design and analysis methods has resulted in a hybrid solution capable of producing energy at a cost between €0.22/kWh and €0.31/kWh depending on the source of funding for the project. Further optimisation through detailed design is expected to lower this further. This thesis develops new and existing methods of design and analysis of wind and wave energy devices. This streamlines the process of early stage development, while adhering to the widely adopted Concept Development Protocol, to develop a technically and economically feasible, combined wind-wave energy hybrid solution.

High-performance 2 mm gate width GaNHEMTs on 6H-SiC with output power of 22.4 W @ 8 GHz

Optimized AlGaN/AlN/GaN high electron mobility transistors (HEMTs) structures were grown on 2-in semi-insulating (SI) 6H-SiC substrate by metal-organic chemical vapor deposition (MOCVD). The 2-in. HEMT wafer exhibited a low average sheet resistance of 305.3 Omega/sq with a uniformity of 3.85%. The fabricated large periphery device with a dimension of 0.35 pm x 2 nun demonstrated high performance, with a maximum DC current density of 1360 mA/mm, a transconductance of 460 mS/mm, a breakdown voltage larger than 80 V, a current gain cut-off frequency of 24 GHz and a maximum oscillation frequency of 34 GHz. Under the condition of continuous-wave (CW) at 9 GHz, the device achieved 18.1 W output power with a power density of 9.05 W/mm and power-added-efficiency (PAE) of 36.4%. While the corresponding results of pulse condition at 8 GHz are 22.4 W output power with 11.2 W/mm power density and 45.3% PAE. These are the state-of-the-art power performance ever reported for this physical dimension of GaN HEMTs based on SiC substrate at 8 GHz. (c) 2008 Elsevier Ltd. All rights reserved.

Power optimization of random distributed feedback fiber lasers

We present a comprehensive study of power output characteristics of random distributed feedback Raman fiber lasers. The calculated optimal slope efficiency of the backward wave generation in the one-arm configuration is shown to be as high as ∼90% for 1 W threshold. Nevertheless, in real applications a presence of a small reflection at fiber ends can appreciably deteriorate the power performance. The developed numerical model well describes the experimental data. © 2012 Optical Society of America.

Experimental Investigation of Performance of a Miniature Direct Methanol Fuel Cell in Short-Term Microgravity

Experimental study of a liquid fed direct methanol fuel cell has been conducted in different gravity environments. A small single cell with 5 cm x 5 cm active area has single serpentine channel on the graphite cathode polar plate and 11 parallel straight channels on the graphite anode flow bed. Cell voltage and current have been measured and two-phase flow in anode channels has been in situ visually observed. The experimental results indicate that the effect of gravity on power performance of the direct methanol fuel cell is large when the concentration polarization governs fuel cells operation. Gravitational effect becomes larger at higher current density. Increasing methanol feeding molarity is conducive to weaken the influence of gravity on performance of liquid fed direct methanol fuel cells. Increasing feeding flow rate of methanol solution from 6 to 15 ml/min could reduce the size of carbon dioxide bubbles, while the influence of gravity still exist. Transport phenomena inside direct methanol fuel cells in microgravity is also analyzed and discussed.

Effect of activated carbon xerogel pore size on the capacitance performance of ionic liquid electrolytes

The use of ionic liquid (IL) electrolytes promises to improve the energy density of electrochemical capacitors (ECs) by allowing for operation at higher voltages. Several studies have also shown that the pore size distribution of materials used to produce electrodes is an important factor in determining EC performance. In this research the capacitative, energy and power performance of ILs 1-ethyl-3- methylimidazolium tetrafluoroborate (EMImBF4), 1-ethyl-3-methylimidazolium dicyanamide (EMImN(CN)2), 1,2-dimethyl-3-propylimidazolium bis(trifluoromethylsulfonyl)imide (DMPImTFSI), and 1-butyl-3-methylpyrrolidinium tris(pentafluoroethyl)trifluorophosphate (BMPyT(F5Et)PF3) were studied and compared with the commercially utilised organic electrolyte 1M tetraethylammonium tetrafluoroborate solution in anhydrous propylene carbonate (Et4NBF4–PC 1 M). To assess the effect of pore size on IL performance, controlled porosity carbons were produced from phenolic resins activated in CO2. The carbon samples were characterised by nitrogen adsorption– desorption at 77 K and the relevant electrochemical behaviour was characterised by cyclic voltammetry, galvanostatic charge–discharge and electrochemical impedance spectroscopy. The best capacitance performance was obtained for the activated carbon xerogel with average pore diameter 3.5 nm, whereas the optimum rate performance was obtained for the activated carbon xerogel with average pore diameter 6 nm. When combined in an EC with IL electrolyte EMImBF4 a specific capacitance of 210 F g1 was obtained for activated carbon sample with average pore diameter 3.5 nm at an operating voltage of 3 V. The activated carbon sample with average pore diameter 6 nm allowed for maximum capacitance retention of approximately 70% at 64 mA cm2.