Computational Power Consumption Research Articles

The Impact of Built-in Ad-Blockers in Web Browsers on Computer Power Consumption

This study investigates the power consumption of various web browsers, specifically focusing on those with built-in ad-blockers compared to standard browsing without ad-blocking features. Using detailed measurements of CPU and GPU power consumption across multiple browsers i.e. Chrome without Ad blocker, Brave, Opera, Firefox, Vivaldi, Librewolf, and Tor—this research highlights the significant impact of ad-blocking on power consumption during web browsing. Experiments were conducted on different types of websites, including video-heavy, news, and entertainment sites, to evaluate how browser optimizations affect overall power usage. Results indicate that browsers with integrated ad-blockers, such as Brave and Librewolf, use significantly reduce power consumption up to 44% compared to traditional browsing setups. The findings also reveal that video content significantly increases CPU and GPU load, with ad-blocking browsers demonstrating superior performance in minimizing energy use. This study emphasizes the importance of browser selection in reducing power consumption, particularly for mobile and battery-dependent devices, and suggests that adopting ad-blocking technologies can lead to substantial energy savings.

Advancements in Nanowire-Based Devices for Neuromorphic Computing: A Review.

Neuromorphic computing, inspired by the highly interconnected and energy-efficient way the human brain processes information, has emerged as a promising technology for post-Moore's law era. This emerging technology can emulate the structures and the functions of the human brain and is expected to overcome the fundamental limitation of the current von Neumann computing architecture. Neuromorphic devices stand out as the key components of future electronic systems, exhibiting potential in shaping the landscape of neuromorphic computing. Especially, nanowire (NW)-based neuromorphic devices, with their advantages of high integration, high-speed computing, and low power consumption, have recently emerged as candidates for neuromorphic computing technology. Here, a critical overview of the current development and relevant research in the field of NW-based neuromorphic devices are provided. Neuromorphic devices based on different NW materials are comprehensively discussed, including Ag NW-based, organic NW-based, metal oxide NW-based, and semiconductor NW-based devices. Finally, as a foresight perspective, the potentials and the challenges of these NW-based neuromorphic devices for use as future brain-like electronics are discussed.

Performance evaluation of resin wafer electrodeionization for enhanced nitrogen recovery from ammonium-containing water environment

Ammonium ion (NH4+) is a common pollutant from various sources like human and animal waste, industrial effluents, agricultural runoff, and commercial products, with traditional treatment methods often generating secondary contaminants, such as nitrite and nitrate. In this study, an energy-efficient resin-wafer electrodeionization (RW-EDI) process was employed for nitrogen recovery from ammonium-rich water. The system was operated in a continuous mode, with voltage-current curve measurements identifying a limiting current density of 2.2 A cm−2. With the NH4+ concentration of 1 g/L and a flow rate of 2.4 mL min−1, the system was tested at various cell voltages (1.2–7.2 V). The system could achieve 80 % NH4+ removal efficiency at a cell voltage of 3 V, current density of 2.2 A cm−2, influent NH4+ level of 3 g/L, and flow rate of 6.4 mL min−1. The observed current efficiency and the calculated power consumption were 41 % and 0.78 kWh m−3, respectively. Finally, a mass balance model was developed to predict the number of RW-EDI stacks for complete removal of NH4+ from the feed. Overall, the result from this study demonstrates that the RW-EDI-based system could be a promising technology for recovering nitrogen from wastewater.

Quantized neural network for complex hologram generation

Computer-generated holography (CGH) is a promising technology for augmented reality displays, such as head-mounted or head-up displays. However, its high computational demand makes it impractical for implementation. Recent efforts to integrate neural networks into CGH have successfully accelerated computing speed, demonstrating the potential to overcome the trade-off between computational cost and image quality. Nevertheless, deploying neural-network-based CGH algorithms on computationally limited embedded systems requires more efficient models with lower computational cost, memory footprint, and power consumption. In this study, we developed a lightweight model for complex hologram generation by introducing neural network quantization. Specifically, we built a model based on tensor holography and quantized it from 32-bit floating-point precision (FP32) to 8-bit integer precision (INT8). Our performance evaluation shows that the proposed INT8 model achieves hologram quality comparable to that of the FP32 model while reducing the model size by approximately 70% and increasing the speed fourfold. Additionally, we implemented the INT8 model on a system-on-module to demonstrate its deployability on embedded platforms and high power efficiency.

Time-series water prediction based on feature clustering fusion and XGboost

Abstract. The time series prediction problem has a very wide range of applications in many fields. Most scholars use the LSTM class of algorithms for prediction. However, this method consumes a significant amount of computational power and presents several challenges. To address this issue, this paper proposes a time series forecasting method based on feature fusion and XGBoost. Specifically, we first utilize holiday information and the K-Means algorithm for feature extraction to expand the feature dimensions of the dataset, and then employ XGBoost as a model for training and prediction. Experiments demonstrate that the method proposed in this paper significantly reduces error compared to other traditional machine learning and deep learning methods, while the training time is much shorter than these methods. For example, compared with LSTM, the MSLE of this model decreases by 1.42%, while the training time is only 0.15% of that of LSTM. This greatly saves on training costs and computational power consumption. This confirms the effectiveness of using machine learning and clustering algorithms in time series prediction and provides new methods and practical application directions for future time series prediction models.

An Efficient RI-MP2 Algorithm for Distributed Many-GPU Architectures.

Second-order Møller-Plesset perturbation theory (MP2) using the Resolution of the Identity approximation (RI-MP2) is a widely used method for computing molecular energies beyond the Hartree-Fock mean-field approximation. However, its high computational cost and lack of efficient algorithms for modern supercomputing architectures limit its applicability to large molecules. In this paper, we present the first distributed-memory many-GPU RI-MP2 algorithm explicitly designed to utilize hundreds of GPU accelerators for every step of the computation. Our novel algorithm achieves near-peak performance on GPU-based supercomputers through the development of a distributed memory algorithm for forming RI-MP2 intermediate tensors with zero internode communication, except for a single asynchronous broadcast, and a distributed memory algorithm for the energy reduction step, capable of sustaining near-peak performance on clusters with several hundred GPUs. Comparative analysis shows our implementation outperforms state-of-the-art quantum chemistry software by over 3.5 times in speed while achieving an 8-fold reduction in computational power consumption. Benchmarking on the Perlmutter supercomputer, our algorithm achieves 11.8 PFLOP/s (83% of peak performance) performing and the RI-MP2 energy calculation on a 314-water cluster with 7850 primary and 30,144 auxiliary basis functions in 4 min on 180 nodes and 720 A100 GPUs. This performance represents a substantial improvement over traditional CPU-based methods, demonstrating significant time-to-solution and power consumption benefits of leveraging modern GPU-accelerated computing environments for quantum chemistry calculations.

Impact of Ad Blockers on Computer Power Consumption while Web Browsing: A Comparative Analysis

This study explores the impact of various ad blockers on power consumption during web browsing, focusing on different types of online content. By analyzing power use across ten popular websites, the study assesses the performance of five widely utilized ad blockers: AdBlock, AdBlock Plus, uBlock, uBlock Origin, and uBlock Origin Lite. Power consumption was measured under controlled conditions, comparing scenarios with and without ad blockers to gain insight into their efficiency. The findings indicate substantial differences in power savings, with some ad blockers significantly reducing power usage, particularly on media-heavy sites, while others unexpectedly increased consumption under certain conditions. The study underscores the potential of ad blockers to enhance power efficiency in digital environments, highlighting the importance of optimizing ad-blocking techniques to reduce the environmental impact of online activities. Through comprehensive analysis and comparison, this research offers insights into selecting effective ad blockers to minimize power consumption, promoting more sustainable web browsing practices.

LB-CLAS: Lattice-based conditional privacy-preserving certificateless aggregate signature scheme for VANET

The rapid development of vehicular ad-hoc networks (VANETs) has brought great convenience to intelligent transportation, and the secure transmission of information in VANETs has become a serious problem. In addition, the protection of private information of vehicles is also a key issue. Aiming at the problem of how to guarantee the secure transmission of information in VANETs under the condition of satisfying security and privacy, we propose a lattice-based conditional privacy-preserving certificateless aggregate signature scheme (LB-CLAS) for VANETs. Instead of using Number Theory Research Unit (NTRU) lattices and discrete Gaussian sampling, the proposed LB-CLAS scheme is based on algebraic lattice. In addition, based on the module version of Small Integer Solution (MSIS) and module version of Learning With Error (MLWE) hard problems, we prove that the LB-CLAS scheme is existential unforgeability under adaptively chosen message attacks (EUF-CMA). Our LB-CLAS scheme employs individual signature verification in vehicle-to-vehicle (V2V) mode, while utilizing aggregate signatures and batch verification in vehicle-to-infrastructure (V2I) mode, with slightly differing transmission parameters between the two modes. Based on Dilithium, our LB-CLAS scheme solves the problem of high storage overhead and computational cost of existing schemes. The performance analysis shows that our LB-CLAS scheme is more efficient in terms of computation cost, storage overhead, and power consumption compared to existing schemes. Compared with existing schemes, our LB-CLAS scheme reduces the signature and verification overheads by more than 17.6% and 43.4%, respectively. Our LB-CLAS program also has significant advantages in batch verification. As the number of vehicles increases, our batch certification time cost is reduced by more than 90%. In addition, our LB-CLAS scheme has the smallest signature length, with a signature size that is 1X smaller than the most efficient existing scheme for the same level of security.

Research on low-power driving fatigue monitoring method based on spiking neural network.

Fatigue driving is one of the leading causes of traffic accidents, and the rapid and accurate detection of driver fatigue is of paramount importance for enhancing road safety. However, the application of deep learning models in fatigue driving detection has long been constrained by high computational costs and power consumption. To address this issue, this study proposes an approach that combines Self-Organizing Map (SOM) and Spiking Neural Networks (SNN) to develop a low-power model capable of accurately recognizing the driver's mental state. Initially, spatial features are extracted from electroencephalogram (EEG) signals using the SOM network. Subsequently, the extracted weight vectors are encoded and fed into the SNN for fatigue driving classification. The research results demonstrate that the proposed method effectively considers the spatiotemporal characteristics of EEG signals, achieving efficient fatigue detection. Simultaneously, this approach successfully reduces the model's power consumption. When compared to traditional artificial neural networks, our method reduces energy consumption by approximately 12.21-42.59%.

IoT-driven optimization of a NxN enhanced pipeline multiplier

As the Internet of Things (IoT) continues to evolve, the demand for efficient hardware solutions is crucial. The paper explores the significance of hardware multiplication in IoT devices and discusses the challenges and opportunities in implementing dedicated hardware accelerators for multiplication operations. It specifically focuses on a highly efficient NxN-bit folded pipeline multiplier designed using advanced 22 nm CMOS technology and a modified hierarchical design approach. This multiplier achieves remarkable performance enhancements by reducing critical path delay through pipelining, resulting in a three-fold increase in operational speed compared to conventional multipliers. Additionally, it consumes only 0.2 mW of power and reduces chip area consumption by over 28 %. The paper emphasizes the importance of minimizing computational latency and power consumption to enable IoT devices to execute complex algorithms with improved speed and resource utilization. It also addresses trade-offs, such as area overhead and design flexibility, associated with hardware multiplication. The insights provided contribute to ongoing efforts to optimize hardware components for IoT applications, fostering the development of more capable and responsive IoT ecosystems. The study presents a rapid and energy-efficient alternative for multiplication in digital hardware, addressing the need for high-performance and energy-efficient devices in the context of the IoT.

Efficient fruit disease diagnosis on resource-constrained agriculture devices

Climate change in recent years has caught the attention of researchers worldwide, with drastic effects on many economic sectors, especially agriculture and livestock. Due to global warming and declining water levels, food security is becoming an increasingly urgent concern. The situation is more challenging in underdeveloped nations such as Pakistan, where little attention has been paid to improving the processes to produce effective seeds and herbicides. The problem has been exacerbated due to financial constraints and a lack of technology to monitor crops. It has been observed that even a noncritical disease can ruin the crop if timely inspections aren’t performed. Although many solutions have been proposed recently, most are based on heavy mobile applications and cloud-based solutions, which results in considerable computational cost, power consumption, and latency. This work proposes a Deep Neural Networks (DNNs)-based light-weight solution for diagnosing and classifying apple crop diseases for farmers working in rural places (where the internet is not accessible). With the increasing need for efficient and robust Machine Learning (ML) applications on edge devices, such as mobile phones, Raspberry Pi, and Jetson Nano-based devices, there has been a demand for more efficient Deep Neural Networks (DNN) models. To meet this demand, various DNN-based models have been experimented with, including Basic CNN Architecture, AlexNet, and EfficientNet Lite to measure performance After a thorough examination and evaluation of each model’s performance, efficiency, and resource usage, the most optimal model was picked and developed applications for classifying apple crop disease. Using a transfer learning strategy on a specially developed EfficientNet DNN architecture, achieving 85% test accuracy.

Analysis, optimization, and testing of a tilt-body aircraft

This paper proposes a new multidisciplinary design optimization framework for tilt-bodies to obtain the optimal combination of geometry and powertrain, achieving both long endurance and low transition power consumption at a 5 kg takeoff weight. A low-fidelity vortex lattice method is used to derive the aerodynamic coefficients. The tandem wings are trimmed using an optimization-based method, considering the effect of distributed propulsion on the trim solution through aero-propulsive interaction at low flight speed. Detailed powertrain parameters are fitted with fewer design variables to calculate power consumption. By analyzing the unique constraints of tilt-bodies, the optimization problem is established as a simultaneous analysis and design architecture, and solved with different numbers of rotors. The impact of the relative geometric relationship between the front and rear wings on aerodynamic performance and constraints is analyzed. Results demonstrate the full angle-of-attack flight capability of tilt-bodies. The optimal design can save nearly 1000 W in transition, with a pure cruise endurance of over 1 hour, which is competitive among other configurations. With an empty weight of only 2 kg, a portion of the battery can be allocated to a heavier payload. High-fidelity computational fluid dynamics corrects the errors of the vortex lattice method on non-lifting components, including the fuselage, nacelles, and landing gear. Finally, the feasibility of the optimal design and the accuracy of the framework are validated by wind tunnel tests and in-flight drag measurements on a test platform.

LSDN Empowers the Metaverse Communication to Achieve High-Resolution Visuals with Lower Bandwidth

Deploying super-resolution models on metaverse terminal devices can enhance visual effects without increasing network bandwidth. However, deploying most current super-resolution networks on metaverse terminal devices with limited hardware resources poses a challenge due to their large volumes and high computing power consumption. In this paper, we present a lightweight separation and distillation network (LSDN) aimed at reducing the model complexity by prioritizing network structure. Specifically, we initially present the blueprint separable convolution (BSConv) to decrease model complexity and utilize the BSConv and information distillation mechanism building the channel separation distillation block (CSDB). Subsequently, we develop the enhanced spatial attention block (ESA) and Fused-MBConv (FMBConv) to explore latent information. In addition, we employ three CSDBs, an ESA, and an FMBConv to construct the residual attention unit (RAU). Finally, we concatenate several RAUs and amalgamate their hierarchical results, and transmit them to the upsampler for reconstructing the high-resolution images. We carried out comprehensive experiments on a range of datasets and found conclusive evidence that the LSDN outperforms state-of-the-art approaches, exhibiting notable enhancements in quantitative and qualitative terms.

Optical cross-correlation analog computing based on dielectric metasurface at mid-infrared wavelengths

Compared to inefficient, energy-intensive digital computing, optical analog computing has significant advantages in terms of computational speed and power consumption. Traditional optical analog computing is typically based on Fourier lenses and spatial frequency filters, which are bulky and difficult to integrate. These disadvantages can be overcome by replacing Fourier lenses and spatial frequency filters in traditional optical analog computing via a metasurface. We showcase a real-time compact optical cross-correlation analog computing based on a dielectric metasurface. An extra phase profile is introduced in the metasurface to approximate the Fresnel diffraction integral as the Fourier transform to realize a more compact scheme. The optical cross-correlation analog computing is derived by directly modulating the impinging optical field. A reference circular pattern of a circular target is highlighted recognized and identified by the operation. The recognition accuracy was analyzed by comparing the recognition peaks with the shifted center of the circle, showing that the operation is affected by the side flap of the other circle pattern in the background. Optical analog computing with a compact and powerful kernel may find potential solutions for bio-detection, pattern recognition, information retrieval, and electron tomography.

Comparing Power Models for circuits design with Mathematical Language Processing

The work presented in this paper is part of a project that focuses on sorting simulation models used at Intel to calculate power consumption of electronic devices. Hundreds of thousands of analytical power models have already been accumulated in a directory of files and folders, for the simulation of the consumption of thousands of products. As a first step, this work focuses on comparing and grouping together models of smaller granularity - the formulas - which consist of power equations, the definition of the parameters involved in these equations, and a description.To manage this large data problem, we follow the principles of Model Management and Analytics, to decompose the problem into the analysis of constituent fields. We believe that the equations describing power consumption are the most highly information-rich field of our models. The aim of this work is to group together the formulas models, using Mathematical Language Processing (MLP), to take advantage of the business information contained in electrical equations. We calculate the embeddings of our formulas, to enable comparisons between them, to cluster together similar equations (and therefore models).

High-Throughput Polar Code Decoders with Information Bottleneck Quantization.

In digital baseband processing, the forward error correction (FEC) unit belongs to the most demanding components in terms of computational complexity and power consumption. Hence, efficient implementation of FEC decoders is crucial for next-generation mobile broadband standards and an ongoing research topic. Quantization has a significant impact on the decoder area, power consumption and throughput. Thus, lower bit widths are preferred for efficient implementations but degrade the error correction capability. To address this issue, a non-uniform quantization based on the Information Bottleneck (IB) method is proposed that enables a low bit width while maintaining the essential information. Many investigations on the use of the IB method for Low-density parity-check code) LDPC decoders exist and have shown its advantages from an implementation perspective. However, for polar code decoder implementations, there exists only one publication that is not based on the state-of-the-art Fast Simplified Successive-Cancellation (Fast-SSC) decoding algorithm, and only synthesis implementation results without energy estimation are shown. In contrast, our paper presents several optimized Fast-SSC polar code decoder implementations using IB-based quantization with placement and routing results using advanced 12 nm FinFET technology. Gains of up to 16% in area and 13% in energy efficiency are achieved with IB-based quantization at a Frame Error Rate (FER) of 10-7 and a polar code of N=1024,R=0.5 compared to state-of-the-art decoders.

Assessment of energy self-sufficiency of a smart farm through integrated modeling of air-source heat pumps and solar power generation

This study explored methods to operate a sustainable system by improving energy independence and integrating renewable energy into a microgrid. However, there are challenges associated with the intermittency of renewable energy and difficulties in predicting the energy demand of energy systems. Therefore, this study emphasizes the need to analyze solar power generation and predict the cooling and heating demands of energy systems. The target system was a smart farm in which hot and cold water was supplied to the smart farm using a heat pump during seasonal changes. A prediction model for the smart farm was developed to calculate power consumption according to the ambient temperature in summer and winter. The results suggested that additional consideration is needed to optimize system operation and renewable energy use to improve the balance between energy consumption and supply. A comparison of the total power consumed to solar energy generation highlighted the challenge of attaining 100% self-sufficiency rates, reaching 44% in summer and 40% in winter. Analysis of solar power generation and air-source heat pump usage trends provided insights into strategies for achieving energy independence in smart farms. In conclusion, this study provides a model for managing energy demands on smart farms and insights into the optimization of system operation with renewable energy. It also provides a foundational framework for evaluating microgrid systems based on renewable energy and contributes to identifying methods to increase the energy self-sufficiency rates to enable sustainability of smart farms.

FPGA-based stereo matching for crop height measurement using monocular camera

We have proposed a hardware-accelerated drone to analyze the condition of farmland right then and there; as a first step, we report that the proposed system can take crop height measurements with high accuracy using a monocular camera. The proposed three-dimensional farmland is generated using stereo matching, where a drone with a monocular camera can extend the parallax distance as the length between two positions when taking a ground image. This means that our approach can improve the accuracy of a reconstructed 3D farmland. In addition, toward real-time computation and low power consumption, the proposed hardware design accelerates image processing efficiently. Thus, to achieve this, we propose a strategy that combines the semi-global matching (SGM) with single path direction and a sum of absolute difference (SAD) with reduced disparity searching length. For example, a semi-global matching (SGM) was employed to smooth the disparity map result before checking the consistency, where the scan line was performed in one direction, from left to right, to speed up the computation time. The experimental result shows that the computation time performed by Xilinx Zynq ZCU102 FPGA achieves 0.77 s for the stereo data set images with 1536 × 1024 pixels resolution. To meet the real-time application and reduce the FPGA resources toward lower power consumption, the experiment discusses reducing the disparity searching length for the SAD computation. In our experiment, the execution time is less than 40 milliseconds, and the circuit volume is around 9,500 LUTs, equivalent to a small-size FPGA. Finally, we also estimated the object's height; a value of 0.43 m was estimated for the object with a physical height of 0.45 m. Meanwhile, for the object with a physical height of 0.65 m, a value of 0.63 m was estimated.

DAM SRAM CORE: An Efficient High-Speed and Low-Power CIM SRAM CORE Design for Feature Extraction Convolutional Layers in Binary Neural Networks.

This article proposes a novel design for an in-memory computing SRAM, the DAM SRAM CORE, which integrates storage and computational functionality within a unified 11T SRAM cell and enables the performance of large-scale parallel Multiply-Accumulate (MAC) operations within the SRAM array. This design not only improves the area efficiency of the individual cells but also realizes a compact layout. A key highlight of this design is its employment of a dynamic aXNOR-based computation mode, which significantly reduces the consumption of both dynamic and static power during the computational process within the array. Additionally, the design innovatively incorporates a self-stabilizing voltage gradient quantization circuit, which enhances the computational accuracy of the overall system. The 64 × 64 bit DAM SRAM CORE in-memory computing core was fabricated using the 55 nm CMOS logic process and validated via simulations. The experimental results show that this core can deliver 5-bit output results with 1-bit input feature data and 1-bit weight data, while maintaining a static power consumption of 0.48 mW/mm2 and a computational power consumption of 11.367 mW/mm2. This showcases its excellent low-power characteristics. Furthermore, the core achieves a data throughput of 109.75 GOPS and exhibits an impressive energy efficiency of 21.95 TOPS/W, which robustly validate the effectiveness and advanced nature of the proposed in-memory computing core design.

Transformer with difference convolutional network for lightweight universal boundary detection.

Although deep-learning methods can achieve human-level performance in boundary detection, their improvements mostly rely on larger models and specific datasets, leading to significant computational power consumption. As a fundamental low-level vision task, a single model with fewer parameters to achieve cross-dataset boundary detection merits further investigation. In this study, a lightweight universal boundary detection method was developed based on convolution and a transformer. The network is called a "transformer with difference convolutional network" (TDCN), which implies the introduction of a difference convolutional network rather than a pure transformer. The TDCN structure consists of three parts: convolution, transformer, and head function. First, a convolution network fused with edge operators is used to extract multiscale difference features. These pixel difference features are then fed to the hierarchical transformer as tokens. Considering the intrinsic characteristics of the boundary detection task, a new boundary-aware self-attention structure was designed in the transformer to provide inductive bias. By incorporating the proposed attention loss function, it introduces the direction of the boundary as strongly supervised information to improve the detection ability of the model. Finally, several head functions with multiscale feature inputs were trained using a bidirectional additive strategy. In the experiments, the proposed method achieved competitive performance on multiple public datasets with fewer model parameters. A single model was obtained to realize universal prediction even for different datasets without retraining, demonstrating the effectiveness of the method. The code is available at https://github.com/neulmc/TDCN.