High CPU Research Articles

A dynamic load balancing model coupled with DAC and ISAT for a stochastic turbulent combustion model

Due to the composition-dependent stiffness of chemistry, simulations of reactive turbulent flows may present computational load imbalance among parallel processes when spatial decomposition is used for parallelisation, causing high CPU idle time and waste of computational resources. To increase computational efficiency, a dynamic load balancing (DLB) model is proposed to redistribute computational load among computing cores. The DLB model exploits a decomposition in the mixture fraction space with two dynamic adjusting decomposition strategies to realise load redistribution. The DLB model is suitable for the integration of chemistry on stochastic particles in hybrid Eulerian/Lagrangian turbulent combustion models in which the Eulerian field is conventionally decomposed statically in physical space in a way that balances the computational load for the solution of the Navier-Stokes equation but which does not generally lead to balanced load for the computation of the composition fields. Here it is tested using an OpenFOAM-based platform, mmcFoam, which is a comprehensive object-orientated C++ library for stochastic turbulent combustion modelling. Apart from direct integration (DI) for chemistry, the DLB model is also coupled with dynamic adaptive chemistry (DAC) and in situ adaptive tabulation (ISAT), which allows for extra speedup. The performance of the coupled models is validated and assessed for two laboratory flame conditions that exhibit different levels of computational load imbalance. Overall, the DLB model effectively balances the computational load distribution and increases the effective usage of computing power, shortening the simulation wall time required. Moreover, a strong scaling test is carried out using up to 512 cores. Although all approaches have sub-ideal scalability, the scalability of each with DLB is significantly better than without DLB. While DLB-ISAT has relatively poor scalability compared to the DI- and DAC-based methods, DLB-ISAT still ranks the fastest among the algorithms in all scaling trials.

Level‐set‐based method for designing novel brushed synchronous machines

AbstractNowadays, research on electromagnetic devices increasingly focuses on multiphysics three‐dimensional complex systems. However, this kind of simulations require huge computational resources and consequently very high CPU time. Ristagno et al. proposed a level‐set based method allowing meshing step savings in any iterative processes (as movement modeling or optimization processes). The level‐set method lies in implicit description of moving fronts as described in the study by Osher et al. (1988). Inspired by Räisänen et al., all physical parameters (material properties, supply or armature movement) are implemented by projection of mathematical functions. In this paper, the authors proposes to study and design a new kind of mechanical commutator for DC machines thanks to this coupled method. Indeed, there is a renewed interest in DC machines because they can be suitable for small mobilities. To be noted, small mobility concerns all kind of vehicles (2 wheels–4 wheels–autonomous vehicles such as drones) that have speeds below 50 km/h. However, specifications for traction applications are difficult to achieve for such a rustic machine and requires some modifications. That is why, inspired by usual synchronous machine, the commutator has been redesigned to obtain expected theoretically signal waveform. Furthermore, numerical coupled simulations have been used to design both machine and supply simultaneously. This work has led to the manufacture of a proof of concept and preliminary results are very promising.

Multi-Class Retinal Diseases Detection Using Deep CNN With Minimal Memory Consumption

Machine Learning (ML) such as Artificial Neural Network (ANN), Deep learning, Recurrent Neural Networks (RNN), Alex Net, and ResNet can be considered as a broad research direction in the identification and classification of critical diseases. CNN and its particular variant, usually named U-Net Segmentation, has made a revolutionary advancement in the classification of medical diseases, specifically retinal diseases. However, because of the feature extraction complexity, U-Net has a significant flaw in high memory and CPU consumption while moving the whole feature map to the corresponding decoder. Furthermore, it can be concatenated to the unsampled decoder feature map avoids reusing pooling indices. In this research work, a convolutional neural network (CNN) model is proposed for multi-class classification problems with the efficient use of memory consumption. The proposed model has been evaluated on a standard benchmark dataset of Eye Net, having 32 classes of retinal diseases. From experimental evaluation, it has been concluded that the proposed model performs better regarding memory management and accuracy. The overall comparison has been performed based on precision, recall, and accuracy with different numbers of epochs and time consumption by each step. The proposed technique achieved an accuracy of 95% on the Eye-net dataset.

PetPS: Supporting Huge Embedding Models with Persistent Memory

Embedding models are effective for learning high-dimensional sparse data. Traditionally, they are deployed in DRAM parameter servers (PS) for online inference access. However, the ever-increasing model capacity makes this practice suffer from both high storage costs and long recovery time. Rapidly developing Persistent Memory (PM) offers new opportunities to PSs owing to its large capacity at low costs, as well as its persistence, while the application of PM also faces two challenges including high read latency and heavy CPU burden. To provide a low-cost but still high-performance parameter service for online inferences, we introduce PetPS, the first production-deployed PM parameter server. (1) To escape with high PM latency, PetPS introduces a PM hash index tailored for embedding model workloads, to minimize PM access. (2) To alleviate the CPU burden, PetPS offloads parameter gathering to NICs, to avoid CPU stalls when accessing parameters on PM and thus improve CPU efficiency. Our evaluation shows that PetPS can boost throughput by 1.3 -- 1.7X compared to PSs that use state-of-the-art PM hash indexes, or get 2.9 -- 5.5X latency reduction with the same throughput. Since 2020, PetPS has been deployed in Kuaishou, one world-leading short video company, and successfully reduced TCO by 30% without performance degradation.

DYNAMIC RESOURCE ALLOCATION ENERGY-EFFICIENT FRAMEWORK FOR GREEN CLOUD COMPUTING

Cloud Computing has been a trending technology for a few years supporting computational services over internet. But ever since its adoption, cloud’s consistent challenge is in its dynamic resource allocation. The existing cloud model details the online and offline algorithms used to decide the dynamic resource allocation. The goal is to have a dynamic resource allocation framework that aligns to cloud data management’s objective of maximizing revenue with minimum cost. This encourages both consumers and cloud providers not only with energy-efficient power usage but also high CPU utilization. This article discusses the impediments of migrating to Public Cloud, what is dynamic resource allocation, HPC workloads with complex communication path on cloud platform, and the benefits of bare metal platform for latency-sensitive applications. We shed light on trade-offs (compute balance) between Private and Public Cloud, how existing resources can be leveraged, Random forest (RF) solutions including a study on hybrid cloud computing capacity optimization framework. Understanding RL architecture, problem solving approach, learning structure and Hybrid Cloud Management Architecture framework are also explored. Also given are a few RL implemented gaming examples on how it makes an impact. Lastly, we shall do the comparisons of RL with other Machine Learning (ML) approaches.

Assessment of LES of intermittent soot production in an aero-engine model combustor using high-speed measurements

Soot production in turbulent flames is an extremely intermittent phenomenon since it is the result of specific thermochemical conditions occasionally occurring in space and time. In realistic configurations such as the swirling flames used in gas-turbines, the presence of large-scale flow motions can additionally affect soot formation processes, leading to even more pronounced intermittency. Classically, the validation of numerical simulations is performed by comparing time-averaged results with experimental data of the phenomenon under investigation. This comparison can be considered as rigorous only if a statistically converged numerical representation is obtained. In case of sporadic events such as intermittent soot formation in turbulent flames, this means to perform the simulation over thousands of milliseconds of physical time, which can have extremely high CPU demands when performing Large Eddy Simulation (LES). In this work, a possible strategy to overcome this issue is proposed based on the use of high-speed measurements and numerically synthesized signals from LES. To illustrate the approach, numerical and experimental soot light scattering signals are considered here by looking at the model aero-engine combustor developed at DLR for the study of pressurized swirled sooting flames. The light scattering signal is numerically synthesized from an LES. Experimental high-speed measurements are used to statistically account for the high temporal and spatial variability of soot when considering time intervals similar to what is today achievable with LES. The feasibility of this approach is finally demonstrated by comparing numerical results to the ensemble of possible soot production states observed experimentally in the DLR burner allowing to eventually validate the present LES results.

ERDAC: Efficient and Reliable Remote Direct Access and Control for Embedded Systems

Emerging embedded systems, such as autonomous vehicles, demand highly efficient remote data transfer, whereas existing networking hardware and protocols cause high communication latency and CPU consumption. In this article, we propose embedded RDAC (eRDAC), an efficient and reliable remote direct access and control solution for embedded systems. The proposed remote access controller in eRDAC has a two-layer protocol offload engine that employs the command/response protocol on UDP to ensure the data reliability and security, and a multichannel DMA controller with configurable priority to improve the efficiency. Besides, a reusable hardware Ethernet MAC is implemented to support not only remote access commands but also standard Ethernet communication. We implement eRDAC on FPGA and the corresponding software in the Linux system. Experimental results show that eRDAC can reduce the latency of remote I/O reading/writing by 74.3%/74.9% ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$3.76\times /3.98\times $ </tex-math></inline-formula> performance improvement) and reduce the latency of remote memory reading and writing with 1024B by 54.2% compared to the socket-based communication. Meanwhile, eRDAC can cut off the consumption of the remote processor and achieve 0.250mJ/Mb energy consumption with only 25-mW power.

SENTunnel: Fast Path for Sensor Data Access on Automotive Embedded Systems

Emerging autonomous vehicles equip multiple high-throughput sensors to enable automatic driving, such as multiline lidars and high-definition cameras. Existing automotive embedded systems usually employ software stacks to receive and preprocess high-throughput sensor data, which brings high latency and CPU consumption. Most research is devoted to accelerators for data processing but ignores the latency overhead caused by sensor data access. Therefore, this article proposes SENTunnel to build fast path from sensors to the corresponding processing units by offloading redundant software stacks into hardware. Specifically, SENTunnel builds fast path for sensor data access to processors/accelerators through two hardware modules. First, the unified access module is used to receive, parse, and transmit raw sensor data. Second, SENTunnel performs necessary preprocessing of different sensor data with the preprocessors module. Based on the design of SENTunnel, we implement a prototype for accessing the data of multiline lidars to the processor and a dedicated accelerator on FPGA. Experimental results indicate that SENTunnel reduces the latency by 55.5% for the data path to processors and reduces the CPU usage caused by the preprocessing driver by 45.9% on average. Compared to the original and partially offloaded data path to accelerators, SENTunnel reduces the latency by 93.8% and 93%, respectively, and eliminates the CPU costs.

Area and power optimization approach for mixed polarity Reed–Muller logic circuits based on multi-strategy bacterial foraging algorithm

Area and power optimization have become the primary constraints in chip design. Existing area and power optimization approaches have poor optimization performance and high CPU time in logic circuits with multiple inputs. To solve this problems, we propose a multi-strategy bacterial foraging algorithm (MBFA) for multi-objective optimization, which includes nondominated add procedure, dual-population cooperative procedure, high-quality replication procedure, distributed migration procedure and the termination criterion. In addition, considering the characteristics of mixed polarity Reed–Muller (MPRM) logic circuits such as high dimension and large solution space domain, we propose an area and power optimization approach (APOA) based on XNOR/OR, which uses MBFA to search MPRM circuits with Pareto optimal solutions. Experimental results show that the APOA has better performance in optimizing MPRM circuits area and power.

Real-time terahertz characterization for composite delamination using a lightweight CPU adaptive network

Delamination defects, as common damages occurring in fiber reinforced composite laminates, are often difficult to be detected due to their invisibility and concealment. Terahertz (THz) technique, as a promising alternative to conventional nondestructive testing (NDT) approaches, has emerged great potentials in quantitative characterization of delamination defects in composites. However, since THz signals are susceptible to confounding interference during THz inspection, such as the noise, interlayer reflection, dispersion and overlap, the accuracy and speed of defect localization and identification are usually contradictory. Generally, to improve the accuracy, many complex manual signal processing methods can be applied, which however will inevitably decrease the inference speed for delamination inspection and cannot fulfill the automatic and real-time analysis requirements in practical THz applications. Here, an intelligent THz 3D characterization system based on the lightweight deep learning (DL) model is proposed to realize the automatic and real-time characterization of hidden delamination defects in composites with high accuracy and low CPU latency. The core aims to specifically design a lightweight and end-to-end CPU adaptive network to obtain the optimal balance between the accuracy and speed. A series of experiments are implemented to validate the superior comprehensive performance of the system in terms of classification accuracy and inference speed on low-cost CPU-based platform. Compared with other methods, the proposed system can be not only able to obtain the superior accuracy, but also to obtain low CPU latency. Therefore, our work provides a novel and general paradigm for DL-based THz intelligent 3D characterization, which will promote the deployment of the system for the automatic and real-time THz characterization in industrial applications.

Acceleration Strategies for MR-STAT: Achieving High-Resolution Reconstructions on a Desktop PC Within 3 Minutes.

MR-STAT is an emerging quantitative magnetic resonance imaging technique which aims at obtaining multi-parametric tissue parameter maps from single short scans. It describes the relationship between the spatial-domain tissue parameters and the time-domain measured signal by using a comprehensive, volumetric forward model. The MR-STAT reconstruction solves a large-scale nonlinear problem, thus is very computationally challenging. In previous work, MR-STAT reconstruction using Cartesian readout data was accelerated by approximating the Hessian matrix with sparse, banded blocks, and can be done on high performance CPU clusters with tens of minutes. In the current work, we propose an accelerated Cartesian MR-STAT algorithm incorporating two different strategies: firstly, a neural network is trained as a fast surrogate to learn the magnetization signal not only in the full time-domain but also in the compressed low-rank domain; secondly, based on the surrogate model, the Cartesian MR-STAT problem is re-formulated and split into smaller sub-problems by the alternating direction method of multipliers. The proposed method substantially reduces the computational requirements for runtime and memory. Simulated and in-vivo balanced MR-STAT experiments show similar reconstruction results using the proposed algorithm compared to the previous sparse Hessian method, and the reconstruction times are at least 40 times shorter. Incorporating sensitivity encoding and regularization terms is straightforward, and allows for better image quality with a negligible increase in reconstruction time. The proposed algorithm could reconstruct both balanced and gradient-spoiled in-vivo data within 3 minutes on a desktop PC, and could thereby facilitate the translation of MR-STAT in clinical settings.

Improvement of cooling of a high heat flux CPU by employing a cooper foam and NEPCM/water suspension

In the current simulation, a copper porous media was inserted on a hot surface of a CPU releasing high heat flux and then embedded within a mini-channel. The working fluid was a mixture of water and Nano-encapsulated phase change material (NEPCM) with fusion temperature of 303.47 K. The proposed model was simulated by computational fluid dynamic (CFD) based on finite volume method (FVM) to evaluate the impacts of Reynolds number (Re = 80, 90, 100, 110, and 120), porous medium with different porosity number (ε = 0.9, 0.925, 0.95, 0.975, and 1). The important parameters in CPU cooling was selected including pressure drop and maximum temperature of CPU's surface. The evidence said that inserting a copper foam with 0.95 porosity can reduce the maximum temperature of CPU's surface and increase the pressure drop by 40 K and 100 times compared to without porous media, respectively. Also, rising Reynolds number from 80 to 120 enhanced the pressure drop by 57 % and decreased the maximum temperature of CPU's surface by only 5.4 K.

Efficient treatment of secondary kinetic processes for pre-partitioned adaptive chemistry approaches

Probability Density Function (PDF) methods, which allow for the direct integration of chemical kinetics, are well established to accurately simulate turbulent flames with strong turbulence-chemistry interactions. While adaptive chemistry techniques have been proven effective in reducing the high CPU cost and memory requirements associated with the handling of chemistry in such simulations, performance metrics have mostly been focussed on the primary oxidation pathways converting fuel to major products. In contrast, this work investigates the ability of adaptive techniques, in this case, the pre-partitioned adaptive chemistry (PPAC) approach, to handle secondary kinetics pathways that are parallel, but tightly coupled to the main oxidation process, taking formation as a case study. PPAC relies on a partitioning of the composition space into a user-specified number of regions, on which specialised reduced models are generated using the Directed Relation Graph with Error Propagation (DRGEP) reduction technique. The direct application of that methodology to a mix of hydrocarbon oxidation and nitrogen-related targets is shown to yield excessively detailed region-specific reduced mechanisms in order to properly capture both the main oxidation and the secondary formation processes, thereby decreasing the benefits of the adaptive approach. To address this issue, a sequential approach is proposed for the generation of the region-specific reduced mechanisms, in which the primary combustion pathways relevant for each region are identified first, followed by the selective addition, directly at the reduced level, of any secondary pathways relevant for that region using a recently developed build-up technique. This new strategy is assessed in the context of propane combustion in a partially stirred reactor (PaSR) and methane combustion in the Sandia Flame D configuration, demonstrating in both cases the benefits of the sequential approach for reduced model generation.

FGLB: A fine‐grained hardware intra‐server load balancer based on 100 G FPGA SmartNIC

AbstractIn today's data centers, workloads including multiple services and requests are processed in parallel within a server with many CPU cores. Therefore, meeting the intra‐server load balancing is very important to improve the utilization of CPU resources in the data centers. However, the existing methods cannot well meet the intra‐server load balancing in high‐throughput scenarios. The software‐based methods generally utilize CPU cores to parse and dispatch packets. They work well at low throughput, but they have high CPU overhead at high throughput, leading to packet loss and high latency issues. The hardware‐based methods parse the packet and compute a hash over its metadata in hardware and perform load balancing in a coarse‐grained manner based on the hash value. They have the ability to work at high throughput with the advantage of low overhead but are less well in balance effect and flexibility. We, therefore, propose an intra‐server load balancer based on the reconfigurable hardware, FPGA, to meet the requirements for load balancing within servers in high‐speed application scenarios. Our method improves the load‐balancing gran‐ ularity of hardware‐based method. It not only has high throughput but also has a good balance effect and flexibility. We implemented and evaluated our method on a 100 G FPGA SmartNIC. The evaluation result shows our method can reduce the load imbalance ratio by an order of magnitude when the distribution of flow size is uneven compared to the current widely used hardware‐based method.

Niyama: Node scheduling for cloud workloads with resource isolation

SummaryCloud providers place tasks from multiple applications on the same resource pool to improve the resource utilization of the infrastructure. The consequent resource contention has an undesirable effect on latency‐sensitive tasks. In this article, we present Niyama—a resource isolation approach that uses a modified version of deadline scheduling to protect latency‐sensitive tasks from CPU bandwidth contention. Conventionally, deadline scheduling has been used to schedule real‐time tasks with well‐defined deadlines. Therefore, it cannot be used directly when the deadlines are unspecified. In Niyama, we estimate deadlines in intervals and secure bandwidth required for the interval, thereby ensuring optimal job response times. We compare our approach with cgroups: Linux's default resource isolation mechanism used in containers today. Our experiments show that Niyama reduces the average delay in tasks by 3–20 when compared to cgroups. Since Linux's deadline scheduling policy is work‐conserving in nature, there is a small drop in the server‐level CPU utilization when Niyama is used naively. We demonstrate how the use of core reservation and oversubscription in the inter‐node scheduler can be used to offset this drop; our experiments show a 1.3–2.24 decrease in delay in job response time over cgroups while achieving high CPU utilization.

RLPlace: Using Reinforcement Learning and Smart Perturbations to Optimize FPGA Placement

Simulated annealing (SA) is one of the most common FPGA placement techniques, and is used both as a standalone algorithm and to improve an initial analytical placement. While SA-based placers can achieve high-quality results, they suffer from long runtimes. In this article, we introduce <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">RLPlace</i> , a novel SA-based FPGA placer that utilizes both reinforcement learning (RL) and targeted perturbations (directed moves). The proposed moves target both wirelength and timing optimization and explore the solution space more efficiently than traditional random moves while preventing oscillation in the Quality of Results (QoR). RL techniques are used to dynamically select the most effective move types as optimization progresses. The experimental results show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">RLPlace</i> outperforms the widely used VTR 8 placer across all runtime/quality tradeoff points, achieving better QoR placement solutions in less runtime. On average, across the Titan23 suite of large FPGA benchmarks, RLPlace can reduce CPU time by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$2.5\times $ </tex-math></inline-formula> with result quality comparable to VTR 8, or improve wirelength by 8% (at a high CPU time budget) −26% (at a low CPU time budget) versus VTR 8.0 given the same CPU time.

Optimum Design of Multi-Section Asymmetrical Transdirectional Couplers with Port Impedance Matching up to S Band Frequency

The method of least squares is developed for the optimum design of multi-section asymmetrical transdirectional (ATRD) couplers with arbitrary port impedance matching. Firstly, transmission matrix of a section of coupler is obtained, then the transmission matrix of whole multi-section coupler is determined, from which the scattering matrix of device is obtained. An error function is then constructed by scattering parameters. According to design specifications, error function depends on physical parameters of coupler (such as widths, spacings and lengths of transmission line sections and capacitor values), of which minimum point gives optimum values of it’s dimensions and capacitors. Optimum design is carried out for a specified frequency response and bandwidth. Error minimization is implemented by the combination of genetic algorithm (GA) and conjugate gradient method (CGM). The design is verified by full-wave simulation software, fabrication and measurement. Several 3 dB, 6 dB and 10 dB couplers are designed with different bandwidths and port impedances. Compared to the proposed design method, design of asymmetrical multi-section couplers by full-wave simulation softwares is by trial and error and at best random and almost impossible because of extremely high CPU time due to the large number of parameters of coupler geometrical structure.

Brain tumor classification by using a novel convolutional neural network structure

AbstractBrain tumors located in the skull are among the health problems that cause serious consequences. Rapid and accurate detection of brain tumor types will ensure that the patient receives appropriate treatment in the early period, thus increasing the patient's chance of recovery and survival. In the literature, classification accuracies over 98% have been acquired automatically by using deep neural networks (DNN) for the brain tumor images such as glioma, meningioma, and pituitary. It is observed that researchers generally focused on achieving higher classification accuracy and therefore, they have used pre‐processing stages, augmentation processes, huge or hybrid DNN structures. These approaches have brought some disadvantages in terms of practical use of the developed methods: (i)The parameters of the pre‐processes should be carefully determined, otherwise the classification accuracy will decrease. (ii) In order to increase the classification performance, it is important to determine the coarse structure of the DNN correctly. If the DNN has many hyper‐parameters, the coarse structure will be determined in a long time. (iii) It is difficult to implement complex DNN structures or training algorithms in terms of practical use, because these methods need huge memory and high CPU computation. In this study, we have proposed a novel DNN model to increase the classification accuracy, and to decrease the number of weights in the structure, and to use less number of hyper‐parameters. We named this model, which uses a divergence‐based feature extractor, as DivFE‐v1 for short. 99.18% classification accuracy for the Figshare dataset is obtained by using the small‐sized DNN structure without any pre‐processing stage or augmentation process.

Multi-source Intelligent Management System of College Snow and Ice Teaching Based on Cloud Platform

The conventional college snow and ice teaching multi-source resource intelligent management system has the problem of incomplete software resource analysis function, which leads to high CPU usage of system. A cloud platform based college snow and ice teaching multi-source resource intelligent management system is designed. Hardware part: record the state of the pins when the register is reset, account for the data bit width and storage capacity of DDR2 SDRAM memory, and optimise the data storage module; Part of the software: obtain ice and snow sports course objectives in colleges and universities, rationally organise the course organization form, optimise the intelligent management mode of teaching multi-source resources by using the cloud platform, support online browsing of various text resources, set relevant parameters to construct various random modification operations, and design the software resource analysis function with the knowledge fusion algorithm. Experimental results: The average CPU usage of the multi-source intelligent resource management system for snow and ice teaching in colleges and universities in this paper and the other two systems is 34.257%, 47.458%, 53.578%. Experimental results show that the performance of multi-source intelligent resource management system for snow and ice teaching in colleges and universities has been significantly improved after making full use of cloud platform.

Generalized Formulation of Response Features for Reliable Optimization of Antenna Input Characteristics

Electromagnetic (EM)-driven parameter adjustment has become imperative in the design of modern antennas. It is necessary because the initial designs rendered through topology evolution, parameter sweeping, or theoretical models are often of poor quality and need to be improved to satisfy stringent performance requirements. Given multiple objectives, constraints, and a typically large number of geometry parameters, the design closure should be carried out through numerical optimization. Unfortunately, standard algorithms entail high CPU expenses and are prone to failure. Feature-based optimization (FBO) is one of the methods developed to alleviate these difficulties by reformulating the design task in terms of the characteristic points extracted from EM-simulated responses. FBO capitalizes on a less nonlinear relationship between the feature point coordinates and antenna dimensions as compared to the original responses (e.g., frequency characteristics). This leads to flattening the functional landscape to be handled, faster convergence of the optimization algorithms, and a possibility of mitigating the issues pertinent to multimodality. Notwithstanding, the response features have to be individually defined for each type of antenna response and tailored to a particular type of design specifications. This requires user experience and hinders the widespread application of FBO. This article proposes a generalized and unified feature point definition, which is suitable for the majority of typical antenna input characteristics (narrowband, multiband, enhanced bandwidth, and wideband) and performance specifications (matching improvement, bandwidth enhancement, and a mixture thereof). Our framework allows for an automated definition of the feature points given the performance specifications, along with their extraction from EM-simulated responses. The operation of the framework is illustrated using a range of planar antennas and favorably compared to conventional (nonfeature-based) design closure task formulation.