Performance Tuning Research Articles

A novel GPU-accelerated strategy for contingency screening of static security analysis

Graphics processing unit (GPU) has been applied successfully in many computation and memory intensive realms due to its superior performances in float-pointing calculation, memory bandwidth and power consumption, and has great potential in power system applications. Contingency screening is a major time consuming part of contingency analysis. In the absence of relevant existing research, this paper is the first of its kind to propose a novel GPU-accelerated algorithm for direct current (DC) contingency screening. Adapting actively unique characteristics of GPU software and hardware, the proposed GPU algorithm is optimized from four aspects: data transmission, parallel task allocation, memory access, and CUDA (Compute Unified Device Architecture) stream. Case studies on a 3012-bus system and 8503-bus system have shown that the GPU-accelerated algorithm, in compared with its counterpart CPU implementation, can achieve about 20 and 50 times speedup respectively. This highly promising performance has demonstrated that carefully designed performance tuning in conjunction with GPU programing architecture is imperative for a GPU-accelerated algorithm. The presented performance tuning strategies can be applicable to other GPU applications in power systems.

Performance assessment and tuning for exchange of clinical documents cross healthcare enterprises

BackgroundTo integrate electronic health records (EHRs) from diverse document sources across healthcare providers, facilities, or medical institutions, the IHE XDS.b profile can be considered as one of the solutions. In this research, we have developed an EHR/OpenXDS system which adopted the OpenXDS, an open source software that complied with the IHE XDS.b profile, and which achieved the EHR interoperability. ObjectiveWe conducted performance testing to investigate the performance and limitations of this EHR/OpenXDS system. MethodologyThe performance testing was conducted for three use cases, EHR submission, query, and retrieval, based on the IHE XDS.b profile for EHR sharing. In addition, we also monitored the depletion of hardware resources (including the CPU usage, memory usage, and network usage) during the test cases execution to detect more details of the EHR/OpenXDS system's limitations. ResultsIn this EHR/OpenXDS system, the maximum affordable workload of the EHR submissions were 400 EHR submissions per hour, the DSA CPU usage was 20%, memory usage was 1380MB, the network usages were 0.286KB input and 7.58KB output per minute; the DPA CPU usage was 1%, memory usage was 1770MB, the network usages were 7.75KB input and 1.54KB output per minute; the DGA CPU usage was 24%, memory usage was 2130MB, the network usages were 1.3KB input and 0.174KB output per minute. The maximum affordable workload of the EHR queries were 600 EHR queries per hour, the DCA CPU usage was 66%, the memory usage was 1660MB, the network usages were 0.230KB input and 0.251KB output per minute; the DGA CPU usage was 1%, the memory usage was 1890MB, the network usages were 0.273KB input and 0.22KB output per minute. The maximum affordable workload of the EHR retrievals were 2000 EHR retrievals, the DCA CPU usage was 79%, the memory usage was 1730MB, the network usages were 19.55KB input and 1.12KB output per minute; the DPA CPU usage was 3.75%, the memory usage was 2310MB, and the network usages were 0.956KB input and 19.57KB output per minute. Discussion and conclusionFrom the research results, we suggest that future implementers who deployed the EHR/OpenXDS system should consider the following aspects. First, to ensure how many service volumes would be provided in the environment and then to adjust the hardware resources. Second, the IHE XDS.b profile is adopted by the SOAP (Simple Object Access Protocol) web service, it might then move onto the Restful (representational state transfer) web service which is more efficient than the SOAP web service. Third, the concurrency process ability should be added in the OpenXDS source code to improve the hardware usage more efficiently while processing the ITI-42, ITI-18, and ITI-43 transactions. Four, this research suggests that the work should continue on adjusting the memory usage for the modules of the OpenXDS thereby using the memory resource more efficiently, e.g., the memory configuration of the JVM (Java Virtual Machine), Apache Tomcat, and Apache Axis2. Fifth, to consider if the hardware monitoring would be required in the implementing environment. These research results provided some test figures to refer to, and it also gave some tuning suggestions and future works to continue improving the performance of the OpenXDS.

Reflective gradient metasurfaces for polarization-independent light focusing at normal or oblique incidence

Reflective gradient metasurfaces are reported as flat, ultra-thin light focusers using a cross-resonator array with spatially varied geometric parameters atop a continuous gold ground plane spaced by a layer of SiO2. The sub-wavelength cross-shaped building element offers polarization-independent performance and full 2π phase tuning range by varying its width and length, which is explained by an analytical model based on harmonically oscillating dipole antenna. With a radial phase gradient, a metasurface is demonstrated to function as a parabolic reflector at 1.47 μm wavelength with the measured efficiency of 44%. In addition, by elaborately engineering the planar distribution of different building elements, another two focusing reflectors are designed and experimentally verified to anomalously reflect and concentrate light along normal direction but with oblique incident angles of 30° and 60°, respectively.

Accuracy-aware wireless indoor localization: Feasibility and applications

Fingerprint-based indoor localization has attracted extensive research efforts due to its potential for deployment without extensive infrastructure support. However, the accuracies of these different systems vary and it is difficult to compare and evaluate these systems systematically. In this work, we propose a Gaussian process based approach that takes the radio map and the localization algorithm as an input, and outputs the expected accuracy of the localization system. With an efficient error estimation algorithm, many applications such as landmark detection, localization algorithm selection and access point subset selection can be performed. Our evaluations show that our approach provides sufficient accuracy and can serve as a useful tool for system evaluation and performance tuning when developing fingerprint-based indoor localization systems.

Performance Tuning and Optimization Techniques of Fixed and Variable Size Batched Cholesky Factorization on GPUs

Solving a large number of relatively small linear systems has recently drawn more attention in the HPC community, due to the importance of such computational workloads in many scientific applications, including sparse multifrontal solvers. Modern hardware accelerators and their architecture require a set of optimization techniques that are very different from the ones used in solving one relatively large matrix. In order to impose concurrency on such throughput-oriented architectures, a common practice is to batch the solution of these matrices as one task offloaded to the underlying hardware, rather than solving them individually.This paper presents a high performance batched Cholesky factorization on large sets of relatively small matrices using Graphics Processing Units (GPUs), and addresses both fixed and variable size batched problems. We investigate various algorithm designs and optimization techniques, and show that it is essential to combine kernel design with performance tuning in order to achieve the best possible performance. We compare our approaches against state-of-the-art CPU solutions as well as GPU-based solutions using existing libraries, and show that, on a K40c GPU for example, our kernels are more than 2× faster.

Performance Optimization of Spin-Torque Microwave Detectors with Material and Operational Parameters

Sensitivity, bandwidth, and noise equivalent power (NEP) are important indicators of the performance of microwave detectors. The previous reports on spin-torque microwave detectors (STMDs) have proposed various approaches to increase the sensitivity. However, the effects of these methods on the other two indicators remain unclear. In this work, macrospin simulation is developed to evaluate how the performance can be optimized through changing the material (tilt angle of reference-layer magnetization) and operational parameters (the direction of magnetic field and working temperature). The study on the effect of magnetic field reveals that the driving force behind the performance tuning is the effective field and the equilibrium angle between the magnetization of the free layer and that of the reference layer. The material that offers the optimal tilt angle in reference-layer magnetization is determined. The sensitivity can be further increased by changing the direction of the applied magnetic field and the operation temperature. Although the optimized sensitivity is accompanied by a reduction in bandwidth or an increase in NEP, a balance among these performance indicators can be reached through optimal tuning of the corresponding influencing parameters.

Auto-tuning for GPGPU applications using performance and energy model

Performance and Energy models are proposed for GPGPU software applications with high fidelity, which can be used for tuning the software parameters.Auto-tuning framework based on simulated annealing and genetic algorithm is proposed for GPGPU software based on the above performance and energy models such that software can be tuned based on user requirements.This work is the first that combines performance and energy tuning automatically for GPGPU software. Display Omitted The general-purpose graphic processing unit (GPGPU) is a popular accelerator for general applications such as scientific computing because the applications are massively parallel and the significant power of parallel computing inheriting from GPUs. However, distributing workload among the large number of cores as the execution configuration in a GPGPU is currently still a manual trial-and-error process. Programmers try out manually some configurations and might settle for a sub-optimal one leading to poor performance and/or high power consumption. This paper presents an auto-tuning approach for GPGPU applications with the performance and power models. First, a model-based analytic approach for estimating performance and power consumption of kernels is proposed. Second, an auto-tuning framework is proposed for automatically obtaining a near-optimal configuration for a kernel computation. In this work, we formulated that automatically finding an optimal configuration as the constraint optimization and solved it using either simulated annealing (SA) or genetic algorithm (GA). Experiment results show that the fidelity of the proposed models for performance and energy consumption are 0.86 and 0.89, respectively. Further, the optimization algorithms result in a normalized optimality offset of 0.94% and 0.79% for SA and GA, respectively.

Automated Performance Tuning of Data Management Systems with Materializations and Indices

Automated performance tuning of data management systems offer various benefits such as improved performance, declined administration costs, and reduced workloads to database administrators (DBAs). Currently, DBAs tune the performance of database systems with a little help from the database servers. In this paper, we propose a new technique for automated performance tuning of data management systems. Firstly, we show how to use the periods of low workload time for performance improvements in the periods of high workload time. We demonstrate that extensions of a database system with materialised views and indices when a workload is low may contribute to better performance for a successive period of high workload. The paper proposes several online algorithms for continuous processing of estimated database workloads and for the discovery of the best plan for materialised view and index database extensions and of elimination of the extensions that are no longer needed. We present the results of experiments that show how the proposed automated performance tuning technique improves the overall performance of a data management system.

Machine Learning Techniques for Power System Security Assessment

Abstract: Modern electricity grids continue to be vulnerable to large-scale blackouts. As all states leading to large-scale blackouts are unique, there is no algorithm to identify pre-emergency states. Moreover, numerical conventional methods are computationally expensive, which makes it difficult to use for the on-line security assessment. Machine learning techniques with their pattern recognition, learning capabilities and high speed of identifying the potential security boundaries can offer an alternative approach. The purpose of this paper is not to suggest that one particular kind of machine learning technique for security assessment would be more appropriate than others. We start from the premise that almost every method may be useful within some restricted context. Based on this idea, we developed an automated multi-model approach for on-line security assessment. The proposed method allows us to automatically test the different state-of-art techniques in order to find both the best algorithm and its top performance tuning for particular analyzed power system. A case study using the IEEE RTC-96 system demonstrates the effectiveness of the proposed approach.

On the Efficiency of Executing Hydro-environmental Models on Cloud

Optimizing high-performance computing applications requires understanding of both the application and its parallelization approach, the system software stack and the target architecture. Traditionally, performance tuning of parallel applications involves consideration of the underlying machine architecture, including floating point performance, memory hierarchies and bandwidth, interconnect architecture, data placement – among others. The shift to the utility computing model through cloud has created tempting economies of scale across IT and domains, not leaving HPC as an exception as a candidate beneficiary. Nevertheless, the infrastructure abstraction and multi-tenancy inherent to cloud offerings poses great challenges to HPC workloads, requiring a dedicated study of applicability of cloud computing as a viable time-to-solution and efficiency platform. In this paper, we present the evaluation of a widely used hydro-environmental code, EFDC, on a cloud platform. Specifically, we evaluate the target parallel application on Linux containers managed by Docker. Unlike virtualization- based solutions that have been widely used for HPC cloud explorations, containers are more fit-for-purpose, sporting among others native execution and lightweight resource consumption. Many-core capability is provided by the OpenMP library in a hybrid configuration with MPI for cross-node data movement, and we explore the combination of these in the target setup. For the MPI part, the work flow is implemented as a data-parallel execution model, with all processing elements performing the same computation, on different sub-domains with thread-level, fine-grain parallelism provided by OpenMP. Optimizing performance requires consideration of the overheads introduced by the OpenMP paradigm such as thread initialization and synchronization. Features of the application make it an ideal test case for deployment on modern cloud architectures, including that it: 1) is legacy code written in Fortran 77, 2) has an implicit solver requiring non-local communication that poses a challenge to traditional partitioning methods, communication optimization and scaling and, 3) is a legacy code across academia, research organizations, governmental agencies, and consulting firms. These technical and practical considerations make this study a representative assessment of migrating legacy codes from traditional HPC systems to the cloud. We finally discuss challenges that stem from the containerized nature of the platform; the latter forms another novel contribution of this paper.

METHODOLOGY FOR ESTIMATION OF HEAT-STRESSED STATE FOR CYLINDER HEAD OF DIESEL ENGINE WITH AIR COOLING

Performance tuning improvement of modern diesel engines exerts an influence on technical and economic, ecological and resource indices of an internal combustion engine (ICE). In its turn, that requires a complex estimation and improvement of engine indices. As is known, working process and thermal intensity of combustion chamber parts are interconnected between themselves. Changes in design, operating and adjustment parameters of systems and structural elements of internal combustion engines that enable air-fuel mixing and combustions processes stressed state of combustion chamber parts. Joint estimation of operating cycle parameters and strength indices of ICE parts makes it possible to solve a number of problems originating in the process of designing new engine models and further development of the existing ones. The paper provides results of comparative estimation on heat-stressed state of a cylinder head for a high-speed diesel engine Д21A (2Ч10,5/12) while operating a rated power mode. In order to simulate and specify a heat-stressed state for description of boundary heat conductivity conditions the following adjoint “gas – wall” problems have been solved: the first one – for combustion chamber of a diesel engine (an internal problem) – simulation of a working cycle for a diesel engine; the second one – for cooling edges while blowing over them by air flow (an external problem). Calculations have been made in three-dimensional non-stationary presentation within the Cartesian coordinates. In order to simulate a working cycle of the diesel engine a computational grid describing combustion chamber configuration, inlet and outlet channels has been used in the paper. Solution of the external problem for cooling edge surface in case of blowing over them by air flow has permitted to specify temperatures and heat-transfer coefficient and later to improve an accuracy while estimating level of temperatures and stresses of the cylinder head.

Tuning framework for stencil computation in heterogeneous parallel platforms

Image processing and computer vision applications are usually complex in terms of the large amount of processed data and high computation loads. To cope with this, optimization techniques and high-performance hardware platforms are required. Since these applications present many opportunities for parallelism, heterogeneous parallel platforms (HPPs) are an interesting choice, offering a good balance between high computation capabilities and flexibility to handle a large spectrum of application features. Applications such as image filtering and edge detection make extensive use of finite difference method to solve partial derivative equations, which computational pattern is called stencil computation. Stencil computations are known as memory-bound, so that reducing high-latency memory access becomes the biggest challenge to reach high performance. In this paper, we present our methodology as a basis of a performance tuning framework to optimize the implementation of multiple stencil computations on HPPs. Results show that our approach outperforms SDK-based methodologies, improving performance. Moreover, using the proposed approach, the developer has the ability of investigating efficiently the performance of the stencil computations before implementing actual code on the target platforms.

Sequential Structural Control of Open-Framework Nanoparticles Both in Dispersion and in Film for Electrochemical Performance Tuning

Abstract We demonstrated that the sequential structural control of nanoparticles of an open-framework coordination polymer, copper hexacyanoferrate (CuHCF). The structural control has been achieved by adding the component ions not only in dispersion phase but also in the thin film form. The CuHCF nanoparticles (NPs) in the dispersion phase changed chemical composition and enhanced dispersibility with capturing [Fe(CN)6]3− anions. On the other hand, the NPs in the thin film were reversibly turned to the original state by capturing Cu2+ cations. Through the sequential processes, all of the synthesized CuHCF NPs were utilized for the high-quality electrochemical electrode for without electrochemical side reaction. This process will contribute the preparation of various electrochemical devices such as electrochromic devices, secondary batteries and biosensors.

Energy-Adaptive Signal Processing Under Renewable Energy

This paper presents an energy-adaptive performance management technique for the design of embedded signal processing systems powered by renewable energy sources. By jointly considering the non-deterministic characteristics of renewable energy and the unique relationship between signal processing performance and the required energy consumption, a progressive performance tuning approach is developed to dynamically determine an acceptable signal processing performance in accordance with the changing energy level at runtime. Several practical issues such as energy prediction errors and battery capacity are investigated, and their impacts on the proposed technique are evaluated. The proposed technique is applied to a DCT-based image sensing system. Simulation results demonstrate that by adaptively tuning signal processing kernels with renewable energy, significant improvements in time coverage and energy efficiency can be achieved in the presence of unstable harvested energy.

Improving network performance on multicore systems: Impact of core affinities on high throughput flows

Network throughput is scaling-up to higher data rates while end-system processors are scaling-out to multiple cores. In order to optimize high speed data transfer into multicore end-systems, techniques such as network adaptor offloads and performance tuning have received a great deal of attention. Furthermore, several methods of multi-threading the network receive process have been proposed. However, thus far attention has been focused on how to set the tuning parameters and which offloads to select for higher performance, and little has been done to understand why the various parameter settings do (or do not) work. In this paper, we build on previous research to track down the sources of the end-system bottleneck for high-speed TCP flows. We define protocol processing efficiency to be the amount of system resources (such as CPU and cache) used per unit of achieved throughput (in Gbps). The amount of various system resources consumed are measured using low-level system event counters. In a multicore end-system, affinitization, or core binding, is the decision regarding how the various tasks of network receive process including interrupt, network, and application processing are assigned to the different processor cores. We conclude that affinitization has a significant impact on protocol processing efficiency, and that the performance bottleneck of the network receive process changes significantly with different affinitization.

복합 이벤트 스트림 질의 처리 성능 개선을 위한 질의 전처리 도구

A complex event processing system, becoming useful in real life domains, efficiently processes stream of continuous events like sensor data from IoT systems. However, those systems do not work well on some types of queries yet, so that programmers should be careful about that. For instance, they do not sufficiently provide detailed guide to choose efficient queries among the almost same meaning queries. In this paper, we propose an query preprocessing tool for event stream processing systems, which helps programmers by giving them the hints to improve performance whenever their queries fall in any possible bad formats in the performance sense. We expect that our proposed module would be a big help to increases productivity of writing programs where debugging, testing, and performance tuning are not straightforward.

Model‐based MPI‐IO tuning with Periscope tuning framework

SummaryFor many parallel applications, I/O performance is a major bottleneck. MPI‐IO, defined by the MPI forum, can help parallel applications overcome the performance and portability limitations of existing parallel I/O interfaces. Although autotuning has been used to improve the performance of computing kernels, MPI‐IO autotuning has rarely been studied. To automate MPI‐IO performance tuning, we designed and implemented an automatic tuner. The tuner relies on the Periscope tuning framework for transparently passing hints to the MPI‐IO library and for automatically collecting performance data. Unlike computational code, each MPI‐IO function takes a relatively long time to complete. Thus, exhaustively searching through the entire parameter space is impractical. So we developed a performance model that can direct us to shorten the tuning time. Copyright © 2015 John Wiley & Sons, Ltd.

ChemInform Abstract: Some Recent Developments in Surface and Interface Design for Photocatalytic and Electrocatalytic Hybrid Structures

AbstractReview: 68 refs.

Bandwidth-variable optical transport network with adjustable time-modulated and inverse-multiplexed wavelength channels

This paper presents a bandwidth-variable and optical-groomable (BVOG) transport network architecture based on time-modulated optical cross-connects that enables efficient switching of periodic burst-sequenced traffic in addition to wavelength streaming traffic; with an elastic control plane that shapes each sub-wavelength traffic stream at the network edge into a sequence of periodic bursts, and inverse-multiplexes a supra-wavelength traffic stream into a group of multiple wavelength traffic streams and a potential sub-wavelength traffic stream to match the bandwidth requirement. BVOG network enables variable-time-slotted lightpaths with variable-length time slots to meet the bandwidth requirements of sub-wavelength traffic; and inverse-multiplexed lightpath-group (IM-lightpath-group) with variable group size to meet the bandwidth requirements of supra-wavelength traffic. For the problem of "routing and wavelength and time assignment" (RWTA) for sub-wavelength traffic; this paper presents an optimal integer linear programming (ILP) formulation that maximizes total network throughput, and a heuristic algorithm to minimize the blocking probability of future connection requests. For the problem of "inverse-multiplexed routing and wavelength assignment" (IMRWA) for supra-wavelength traffic; this paper presents an optimal ILP formulation that minimizes the total number of hops and the total per-hop propagation delays of the IM-lightpath-group, and a heuristic algorithm that minimizes the total number of wavelength channels of the IM-lightpath-group. The RWTA ILP is solved by the ILOG CPLEX optimization tool, and it is used to check the optimality of the RWTA heuristic. Performance tuning of the IMRWA heuristic is investigated to balance the tradeoff between performance and computation complexity. Performance tuning of the BVOG network with the RWTA and IMRWA heuristic is examined through simulation experiments.

Performance tuning and sparse traversal technique for a cell‐based fetch length algorithm on a GPU

SummaryFor determining distances (fetch lengths) from points to polygons in a two‐dimensional Euclidean plane, cell‐based algorithms provide a simple and effective solution. They divide the input area into a grid of cells that cover the area. The objects are stored into the appropriate cells, and the resulting structure is used for solving the problem. When the input objects are distributed unevenly or the cell size is small, most of the cells may be empty. The representation is then called sparse. In the method proposed in this work, each cell contains information about its distance to the nonempty cells. It is then possible to skip over several empty cells at a time without memory accesses. A cell‐based fetch length algorithm is implemented on a graphics processing unit (GPU). Because control flow divergence reduces its performance, several methods to reduce the divergence are studied. While many of the explicit attempts turn out to be unsuccessful, sorting of the input data and sparse traversal are observed to greatly improve performance: compared with the initial GPU implementation, up to 45‐fold speedup is reached. The speed improvement is greatest when the map is very sparse and the points are given in a random order. Copyright © 2015 John Wiley & Sons, Ltd.