Graphics Processing Units(GPUs)are used to accelerate computing-intensive tasks,such as neural networks,data analysis,high-performance computing,etc.In the past decade or so,researchers have done a lot of work on GPU ...Graphics Processing Units(GPUs)are used to accelerate computing-intensive tasks,such as neural networks,data analysis,high-performance computing,etc.In the past decade or so,researchers have done a lot of work on GPU architecture and proposed a variety of theories and methods to study the microarchitectural characteristics of various GPUs.In this study,the GPU serves as a co-processor and works together with the CPU in an embedded real-time system to handle computationally intensive tasks.It models the architecture of the GPU and further considers it based on some excellent work.The SIMT mechanism and Cache-miss situation provide a more detailed analysis of the GPU architecture.In order to verify the GPU architecture model proposed in this article,10 GPU kernel_task and an Nvidia GPU device were used to perform experiments.The experimental results showed that the minimum error between the kernel task execution time predicted by the GPU architecture model proposed in this article and the actual measured kernel task execution time was 3.80%,and the maximum error was 8.30%.展开更多
As the basis of designing and implementing a cyber-physical system (CPS), architecture research is very important but still at preliminary stage. Since CPS includes physical components, time and space constraints seri...As the basis of designing and implementing a cyber-physical system (CPS), architecture research is very important but still at preliminary stage. Since CPS includes physical components, time and space constraints seriously challenge architecture study. In this paper, a service-oriented architecture of CPS was presented. Further, a two-way time synchronization algorithm for CPS service composition was put forward. And a formal method, for judging if actual CPS service meets space constraints, was suggested, which was based on space-π-calculus proposed. Finally, a case study was performed and CPS business process designed by the model and the proposed methods could run well. The application of research conclusion implies that it has rationality and feasibility.展开更多
With the advent of Big Data, the fields of Statistics and Computer Science coexist in current information systems. In addition to this, technological advances in embedded systems, in particular Internet of Things tech...With the advent of Big Data, the fields of Statistics and Computer Science coexist in current information systems. In addition to this, technological advances in embedded systems, in particular Internet of Things technologies, make it possible to develop real-time applications. These technological developments are disrupting Software Engineering because the use of large amounts of real-time data requires advanced thinking in terms of software architecture. The purpose of this article is to propose an architecture unifying not only Software Engineering and Big Data activities, but also batch and streaming architectures for the exploitation of massive data. This architecture has the advantage of making possible the development of applications and digital services exploiting very large volumes of data in real time;both for management needs and for analytical purposes. This architecture was tested on COVID-19 data as part of the development of an application for real-time monitoring of the evolution of the pandemic in Côte d’Ivoire using PostgreSQL, ELasticsearch, Kafka, Kafka Connect, NiFi, Spark, Node-Red and MoleculerJS to operationalize the architecture.展开更多
Precoding and space-time block coding （STBC）techniques using the uniform channel decomposition （UCD）are proposed to improve the bit error rate （BER） of themultiple-antenna communication system, but at a cost of ...Precoding and space-time block coding （STBC）techniques using the uniform channel decomposition （UCD）are proposed to improve the bit error rate （BER） of themultiple-antenna communication system, but at a cost of areduced data rate. In order to achieve a higher overall systemperformance, a novel adaptive transceiver architecture whichflexibly combines both UCD and UCD ＋ STBC technologies isproposed. The channel state information （CSI） feedback pathwas added to the precoder to select which coding method wasto be used, i.e. UCD alone or UCD ＋ STBC. With thesmaller constellation sizes, Matlab simulation results showthat, the adaptive transceiver architecture will select the UCD-only mode under the higher SNR conditions in order to achievea higher bit rate. The UCD ＋ STBC mode will be selectedunder the lower SNR conditions （e. g., SNR 〈 10 dB） inorder to maintain good BER performance at the cost of areduced data rate. This architecture was implemented andverified using both UMC 0.18 ASIC process technology and aXilinx xc4vlx Virtex-4 FPGA at 150 MHz. The simulationresults demonstrate that the required number of reconfigurablearithmetic unit slices grows linearly with the channel matrixsize, while the number of adder array unit and reconfigurablelogic unit slices increases slightly with the constellation size.展开更多
The concept of resilient positioning, navigation and timing (PNT) is described. The definition of resilient PNT is given, the relationship between integrated (or comprehensive) PNT and resilient PNT is analyzed, and i...The concept of resilient positioning, navigation and timing (PNT) is described. The definition of resilient PNT is given, the relationship between integrated (or comprehensive) PNT and resilient PNT is analyzed, and it is pointed out that the integrated PNT is the foundation of resilient PNT. Resilient PNT should be divided into resilient sensor integration, resilient functional model and resilient stochastic model. The strategy and principles of resilient integration of sensors are discussed. The resilient integration of sensors should be designed following the optimal, available, compatible and interoperable principles. The concepts of resilient functional model and possible modification strategies of the different functional models are also described. Several possible optimal routes for resilient stochastic model improvements are also set forth. It is pointed out that the optimal improvements of stochastic models for multi PNT sources should follow the same variance scale. At last, the resilient PNT data fusion for state parameters are given based on the resilient functional and stochastic models.展开更多
文摘Graphics Processing Units(GPUs)are used to accelerate computing-intensive tasks,such as neural networks,data analysis,high-performance computing,etc.In the past decade or so,researchers have done a lot of work on GPU architecture and proposed a variety of theories and methods to study the microarchitectural characteristics of various GPUs.In this study,the GPU serves as a co-processor and works together with the CPU in an embedded real-time system to handle computationally intensive tasks.It models the architecture of the GPU and further considers it based on some excellent work.The SIMT mechanism and Cache-miss situation provide a more detailed analysis of the GPU architecture.In order to verify the GPU architecture model proposed in this article,10 GPU kernel_task and an Nvidia GPU device were used to perform experiments.The experimental results showed that the minimum error between the kernel task execution time predicted by the GPU architecture model proposed in this article and the actual measured kernel task execution time was 3.80%,and the maximum error was 8.30%.
基金National High-Tech Research and Development Programs of China( 863 Program) ( No. 2011AA010101,No. 2012AA062203) National Natural Science Foundation of China ( No. 61103069 ) Key Research Project of Shanghai Science and Technology Committee,China( No. 10dz1122600)
文摘As the basis of designing and implementing a cyber-physical system (CPS), architecture research is very important but still at preliminary stage. Since CPS includes physical components, time and space constraints seriously challenge architecture study. In this paper, a service-oriented architecture of CPS was presented. Further, a two-way time synchronization algorithm for CPS service composition was put forward. And a formal method, for judging if actual CPS service meets space constraints, was suggested, which was based on space-π-calculus proposed. Finally, a case study was performed and CPS business process designed by the model and the proposed methods could run well. The application of research conclusion implies that it has rationality and feasibility.
文摘With the advent of Big Data, the fields of Statistics and Computer Science coexist in current information systems. In addition to this, technological advances in embedded systems, in particular Internet of Things technologies, make it possible to develop real-time applications. These technological developments are disrupting Software Engineering because the use of large amounts of real-time data requires advanced thinking in terms of software architecture. The purpose of this article is to propose an architecture unifying not only Software Engineering and Big Data activities, but also batch and streaming architectures for the exploitation of massive data. This architecture has the advantage of making possible the development of applications and digital services exploiting very large volumes of data in real time;both for management needs and for analytical purposes. This architecture was tested on COVID-19 data as part of the development of an application for real-time monitoring of the evolution of the pandemic in Côte d’Ivoire using PostgreSQL, ELasticsearch, Kafka, Kafka Connect, NiFi, Spark, Node-Red and MoleculerJS to operationalize the architecture.
基金The National Natural Science Foundation of China(No.61376025)the Industry-Academic Joint Technological Innovations FundP roject of Jiangsu(No.BY2013003-11)the Scientific Innovation Research of College Graduates in Jiangsu Province(No.KYLX_0273)
文摘Precoding and space-time block coding （STBC）techniques using the uniform channel decomposition （UCD）are proposed to improve the bit error rate （BER） of themultiple-antenna communication system, but at a cost of areduced data rate. In order to achieve a higher overall systemperformance, a novel adaptive transceiver architecture whichflexibly combines both UCD and UCD ＋ STBC technologies isproposed. The channel state information （CSI） feedback pathwas added to the precoder to select which coding method wasto be used, i.e. UCD alone or UCD ＋ STBC. With thesmaller constellation sizes, Matlab simulation results showthat, the adaptive transceiver architecture will select the UCD-only mode under the higher SNR conditions in order to achievea higher bit rate. The UCD ＋ STBC mode will be selectedunder the lower SNR conditions （e. g., SNR 〈 10 dB） inorder to maintain good BER performance at the cost of areduced data rate. This architecture was implemented andverified using both UMC 0.18 ASIC process technology and aXilinx xc4vlx Virtex-4 FPGA at 150 MHz. The simulationresults demonstrate that the required number of reconfigurablearithmetic unit slices grows linearly with the channel matrixsize, while the number of adder array unit and reconfigurablelogic unit slices increases slightly with the constellation size.
文摘The concept of resilient positioning, navigation and timing (PNT) is described. The definition of resilient PNT is given, the relationship between integrated (or comprehensive) PNT and resilient PNT is analyzed, and it is pointed out that the integrated PNT is the foundation of resilient PNT. Resilient PNT should be divided into resilient sensor integration, resilient functional model and resilient stochastic model. The strategy and principles of resilient integration of sensors are discussed. The resilient integration of sensors should be designed following the optimal, available, compatible and interoperable principles. The concepts of resilient functional model and possible modification strategies of the different functional models are also described. Several possible optimal routes for resilient stochastic model improvements are also set forth. It is pointed out that the optimal improvements of stochastic models for multi PNT sources should follow the same variance scale. At last, the resilient PNT data fusion for state parameters are given based on the resilient functional and stochastic models.
