On the potential of significance-driven execution for energy-aware HPC

Dynamic Voltage and Frequency Scaling (DVFS) exhibits fundamental limitations as a method to reduce energy consumption in computing systems. In the HPC domain, where performance is of highest priority and codes are heavily optimized to minimize idle time, DVFS has limited opportunity to achieve substantial energy savings. This paper explores if operating processors Near the transistor Threshold Volt- age (NTV) is a better alternative to DVFS for break- ing the power wall in HPC. NTV presents challenges, since it compromises both performance and reliability to reduce power consumption. We present a first of its kind study of a significance-driven execution paradigm that selectively uses NTV and algorithmic error tolerance to reduce energy consumption in performance- constrained HPC environments. Using an iterative algorithm as a use case, we present an adaptive execution scheme that switches between near-threshold execution on many cores and above-threshold execution on one core, as the computational significance of iterations in the algorithm evolves over time. Using this scheme on state-of-the-art hardware, we demonstrate energy savings ranging between 35% to 67%, while compromising neither correctness nor performance.

TProf: An energy profiler for task-parallel programs

We present TProf, an energy profiling tool for OpenMP-like task-parallel programs. To compute the energy consumed by each task in a parallel application, TProf dynamically traces the parallel execution and uses a novel technique to estimate the per-task energy consumption. To achieve this estimation, TProf apportions the total processor energy among cores and overcomes the limitation of current works which would otherwise make parallel accounting impossible to achieve. We demonstrate the value of TProf by characterizing a set of task parallel programs, where we find that data locality, memory access patterns and task working sets are responsible for significant variance in energy consumption between seemingly homogeneous tasks. In addition, we identify opportunities for fine-grain energy optimization by applying per-task Dynamic Voltage and Frequency Scaling (DVFS).

Power-Capped DVFS and Thread Allocation with ANN Models on Modern NUMA Systems

Abstract—Power capping is an essential function for efficient power budgeting and cost management on modern server systems. Contemporary server processors operate under power caps by using dynamic voltage and frequency scaling (DVFS). However, these processors are often deployed in non-uniform memory
access (NUMA) architectures, where thread allocation between cores may significantly affect performance and power consumption. This paper proposes a method which maximizes performance under power caps on NUMA systems by dynamically optimizing two knobs: DVFS and thread allocation. The method selects the optimal combination of the two knobs with models based on artificial neural network (ANN) that captures the nonlinear effect of thread allocation on performance. We implement
the proposed method as a runtime system and evaluate it with twelve multithreaded benchmarks on a real AMD Opteron based NUMA system. The evaluation results show that our method outperforms a naive technique optimizing only DVFS by up to
67.1%, under a power cap.

Distribution Network Voltage Support Using Sensitivity-based Dispatch of Distributed Generation

This paper is concerned with the voltage and reactive power issues surrounding the connection of Distributed Generation (DG) on the low-voltage (LV) distribution network. The presented system-wide voltage control algorithm consists of three stages. Firstly available reactive power reserves are utilized. Then, if required, DG active power output is curtailed. Finally, curtailment of non-critical site demand is considered. The control methodology is tested on a variant of the 13-bus IEEE Node Radial Distribution Test Feeder. The presented control algorithm demonstrated that the distribution system operator (DSO) can maintain voltage levels within a desired statutory range by dispatching reactive power from DG or network devices. The practical application of the control strategy is discussed.

Power Scaling of Uplink Massive MIMO Systems with Arbitrary-Rank Channel Means

This paper investigates the uplink achievable rates of massive multiple-input multiple-output (MIMO) antenna systems in Ricean fading channels, using maximal-ratio combining (MRC) and zero-forcing (ZF) receivers, assuming perfect and imperfect channel state information (CSI). In contrast to previous relevant works, the fast fading MIMO channel matrix is assumed to have an arbitrary-rank deterministic component as well as a Rayleigh-distributed random component. We derive tractable expressions for the achievable uplink rate in the large-antenna limit, along with approximating results that hold for any finite number of antennas. Based on these analytical results, we obtain the scaling law that the users' transmit power should satisfy, while maintaining a desirable quality of service. In particular, it is found that regardless of the Ricean K-factor, in the case of perfect CSI, the approximations converge to the same constant value as the exact results, as the number of base station antennas, M, grows large, while the transmit power of each user can be scaled down proportionally to 1/M. If CSI is estimated with uncertainty, the same result holds true but only when the Ricean K-factor is non-zero. Otherwise, if the channel experiences Rayleigh fading, we can only cut the transmit power of each user proportionally to 1/√M. In addition, we show that with an increasing Ricean K-factor, the uplink rates will converge to fixed values for both MRC and ZF receivers.

Massive MIMO with Non-Ideal Arbitrary Arrays: Hardware Scaling Laws and Circuit-Aware Design

Massive multiple-input multiple-output (MIMO) systems are cellular networks where the base stations (BSs) are equipped with unconventionally many antennas, deployed on colocated or distributed arrays. Huge spatial degrees-of-freedom are achieved by coherent processing over these massive arrays, which provide strong signal gains, resilience to imperfect channel knowledge, and low interference. This comes at the price of more infrastructure; the hardware cost and circuit power consumption scale linearly/affinely with the number of BS antennas N. Hence, the key to cost-efficient deployment of large arrays is low-cost antenna branches with low circuit power, in contrast to today’s conventional expensive and power-hungry BS antenna branches. Such low-cost transceivers are prone to hardware imperfections, but it has been conjectured that the huge degrees-of-freedom would bring robustness to such imperfections. We prove this claim for a generalized uplink system with multiplicative phasedrifts, additive distortion noise, and noise amplification. Specifically, we derive closed-form expressions for the user rates and a scaling law that shows how fast the hardware imperfections can increase with N while maintaining high rates. The connection between this scaling law and the power consumption of different transceiver circuits is rigorously exemplified. This reveals that one can make the circuit power increase as p N, instead of linearly, by careful circuit-aware system design.

VEDA:Variation-aware energy-efficient discrete Wavelet Transform architecture

In this paper, we present a unified approach to an energy-efficient variation-tolerant design of Discrete Wavelet Transform (DWT) in the context of image processing applications. It is to be noted that it is not necessary to produce exactly correct numerical outputs in most image processing applications. We exploit this important feature and propose a design methodology for DWT which shows energy quality tradeoffs at each level of design hierarchy starting from the algorithm level down to the architecture and circuit levels by taking advantage of the limited perceptual ability of the Human Visual System. A unique feature of this design methodology is that it guarantees robustness under process variability and facilitates aggressive voltage over-scaling. Simulation results show significant energy savings (74% - 83%) with minor degradations in output image quality and avert catastrophic failures under process variations compared to a conventional design. © 2010 IEEE.

Analysis of voltage source converter-based high-voltage direct current under DC line-to-earth fault

DC line faults on high-voltage direct current (HVDC) systems utilising voltage source converters (VSCs) are a major issue for multi-terminal HVDC systems in which complete isolation of the faulted system is not a viable option. Of these faults, single line-to-earth faults are the most common fault scenario. To better understand the system under such faults, this study analyses the behaviour of HVDC systems based on both conventional two-level converter and multilevel modular converter technology, experiencing a permanent line-to-earth fault. Operation of the proposed system under two different earthing configurations of converter side AC transformer earthed with converter unearthed, and both converter and AC transformer unearthed, was analysed and simulated, with particular attention paid to the converter operation. It was observed that the development of potential earth loops within the system as a result of DC line-to-earth faults leads to substantial overcurrent and results in oscillations depending on the earthing configuration.

Power Modeling and Capping for Heterogeneous ARM/FPGA SoCs

Low-power processors and accelerators that were originally designed for the embedded systems market are emerging as building blocks for servers. Power capping has been actively explored as a technique to reduce the energy footprint of high-performance processors. The opportunities and limitations of power capping on the new low-power processor and accelerator ecosystem are less understood. This paper presents an efficient power capping and management infrastructure for heterogeneous SoCs based on hybrid ARM/FPGA designs. The infrastructure coordinates dynamic voltage and frequency scaling with task allocation on a customised Linux system for the Xilinx Zynq SoC. We present a compiler-assisted power model to guide voltage and frequency scaling, in conjunction with workload allocation between the ARM cores and the FPGA, under given power caps. The model achieves less than 5% estimation bias to mean power consumption. In an FFT case study, the proposed power capping schemes achieve on average 97.5% of the performance of the optimal execution and match the optimal execution in 87.5% of the cases, while always meeting power constraints.

Fair charging strategies for EVs connected to a low-voltage distribution network

The development of smart grid technologies and appropriate charging strategies are key to accommodating large numbers of Electric Vehicles (EV) charging on the grid. In this paper a general framework is presented for formulating the EV charging optimization problem and three different charging strategies are investigated and compared from the perspective of charging fairness while taking into account power system constraints. Two strategies are based on distributed algorithms, namely, Additive Increase and Multiplicative Decrease (AIMD), and Distributed Price-Feedback (DPF), while the third is an ideal centralized solution used to benchmark performance. The algorithms are evaluated using a simulation of a typical residential low voltage distribution network with 50% EV penetration. © 2013 IEEE.

Investigation of AIMD based charging strategies for EVs connected to a low-voltage distribution network

In this paper we consider charging strategies that mitigate the impact of domestic charging of EVs on low-voltage distribution networks and which seek to reduce peak power by responding to time-ofday pricing. The strategies are based on the distributed Additive Increase and Multiplicative Decrease (AIMD) charging algorithms proposed in [5]. The strategies are evaluated using simulations conducted on a custom OpenDSS-Matlab platform for a typical low voltage residential feeder network. Results show that by using AIMD based smart charging 50% EV penetration can be accommodated on our test network, compared to only 10% with uncontrolled charging, without needing to reinforce existing network infrastructure. © Springer-Verlag Berlin Heidelberg 2013.

On-off based charging strategies for EVs connected to a Low Voltage distributon network

The development of appropriate Electric Vehicle (EV) charging strategies has been identified as an effective way to accommodate an increasing number of EVs on Low Voltage (LV) distribution networks. Most research studies to date assume that future charging facilities will be capable of regulating charge rates continuously, while very few papers consider the more realistic situation of EV chargers that support only on-off charging functionality. In this work, a distributed charging algorithm applicable to on-off based charging systems is presented. Then, a modified version of the algorithm is proposed to incorporate real power system constraints. Both algorithms are compared with uncontrolled and centralized charging strategies from the perspective of both utilities and customers. © 2013 IEEE.

Scaling of magneto-quantum-radiative hydrodynamic equations: From laser-produced plasmas to astrophysics

We introduce the equations of magneto-quantum-radiative hydrodynamics. By rewriting them in a dimensionless form, we obtain a set of parameters that describe scale-dependent ratios of characteristic hydrodynamic quantities. We discuss how these dimensionless parameters relate to the scaling between astrophysical observations and laboratory experiments.

On the scaling behavior of electric conductivity in [C(4)mim][NTf2]

In this work we examine, for the first time, the molar conductivity behavior of the deeply supercooled room temperature ionic liquid [C₄mim][NTf₂] in the temperature, pressure and volume thermodynamic space in terms of density scaling (TV^γ)⁻¹ combined with the equation of state (EOS). The exponent γ_σ determined from the Avramov model analysis is compared with the coefficient obtained from the viscosity studies carried out at moderate temperatures. Therefore, the experimental results presented herein provide the answer to the long-standing question regarding the validity of thermodynamic scaling of ionic liquids over a wide temperature range, i.e. from the normal liquid state to the glass transition point. Finally, we investigate the relationship between the dynamic and thermodynamic properties of [C₄mim][NTf₂] represented by scaling exponent γ and Grüneisen constant γ_G, respectively.

A quality-scalable and energy-efficient approach for spectral analysis of heart rate variability

Today there is a growing interest in the integration of health monitoring applications in portable devices necessitating the development of methods that improve the energy efficiency of such systems. In this paper, we present a systematic approach that enables energy-quality trade-offs in spectral analysis systems for bio-signals, which are useful in monitoring various health conditions as those associated with the heart-rate. To enable such trade-offs, the processed signals are expressed initially in a basis in which significant components that carry most of the relevant information can be easily distinguished from the parts that influence the output to a lesser extent. Such a classification allows the pruning of operations associated with the less significant signal components leading to power savings with minor quality loss since only less useful parts are pruned under the given requirements. To exploit the attributes of the modified spectral analysis system, thresholding rules are determined and adopted at design- and run-time, allowing the static or dynamic pruning of less-useful operations based on the accuracy and energy requirements. The proposed algorithm is implemented on a typical sensor node simulator and results show up-to 82% energy savings when static pruning is combined with voltage and frequency scaling, compared to the conventional algorithm in which such trade-offs were not available. In addition, experiments with numerous cardiac samples of various patients show that such energy savings come with a 4.9% average accuracy loss, which does not affect the system detection capability of sinus-arrhythmia which was used as a test case.