Recent code-based test input generators based on dynamic symbolic execution increase path coverage by solving path condition with a constraint or an SMT solver. When the solver considers path condition produced from a...
详细信息
Recent code-based test input generators based on dynamic symbolic execution increase path coverage by solving path condition with a constraint or an SMT solver. When the solver considers path condition produced from an infeasible path, it tries to show unsatisfiability, which is a useless time-consuming process. In this paper, we propose a new method that takes opportunity of the detection of a single infeasible path to generalize to a (possibly infinite) family of infeasible paths, which will not have to be considered in further path conditions solving. The method exploits non-intrusive constraint-based explanations, a technique developed in Constraint Programming to explain unsatisfiability. Experimental results obtained with our prototype tool IPEG show that, whatever is the underlying constraint solving procedure (IC, Colibri and the SMT solver Z3), this approach can save considerable computational time.
The Munich Simulation Computer (MuSiC), a special-purpose, highly-parallel programmable machine, is an approach to transfer concepts (developed for dataflow computers) to fast, mixed-design-level simulation of digita...
详细信息
The Munich Simulation Computer (MuSiC), a special-purpose, highly-parallel programmable machine, is an approach to transfer concepts (developed for dataflow computers) to fast, mixed-design-level simulation of digital systems. To gain high performance, however, the operation principles is modified from dataflow computation to event flow computation. This paper presents the event flow computation scheme and its implementation by the MUSIC organisation. Parameters, which influence performance, are discussed and it is shown that MuSiC can simulate more than 8 million gates and flipflops at a speed of up to some billion events per second.
This paper presents a new design approach for the inverse discrete sine and cosine transforms based on a new unified systolic array algorithm that can be used to obtain VLSI array architectures with very small differe...
详细信息
This paper presents a new design approach for the inverse discrete sine and cosine transforms based on a new unified systolic array algorithm that can be used to obtain VLSI array architectures with very small differences in the hardware structure for the two transforms and improved performances. This new algorithm is based on using highly regular and modular computational structures as circular correlation and can be mapped into linear systolic arrays with a small number of I/O channels and low I/O bandwidth.
The development of Medical Image Analysis (MIA) applications that can successfully be applied in clinical practice is difficult for several reasons, one of them being the large amount and variety of resources involved...
详细信息
The development of Medical Image Analysis (MIA) applications that can successfully be applied in clinical practice is difficult for several reasons, one of them being the large amount and variety of resources involved (people, data, methods, computing). The application goes through several phases (development, parameter optimization, evaluation and clinical deployment) usually supported by different systems. The lack of support for information flow from phase to phase puts extra logistics burden on the lifecycle of MIA applications. In this paper we describe our first efforts to develop a Problem Solving Environment (PSE) for MIA applications using the three systems available at the proof-of-concept environment of the Virtual Laboratory for e-Sciences project. The proposed PSE implements data provenance mechanisms that support information flow among systems, facilitating navigation across phases of the application lifecycle.
In high performance computing environment in many fields, to finish a task requires not only computing service, but also data service and all kinds of device services, which results in different relation dependences a...
详细信息
In high performance computing environment in many fields, to finish a task requires not only computing service, but also data service and all kinds of device services, which results in different relation dependences and is very complex, however, job management systems such as PBS, LSF, and Condor and so on, only support time dependence among jobs, and are short of fault-tolerance and complete flow control. Workflow-based user environment for high performance computing is successfully applied in railway survey, which not only benefits construction and control of task flow, but also enhances all kinds of relation dependences and flow semantics, especially for the work flow engine based on event model with checkpoint function. It provides good reliability. Thus, workflow-based user environment for high performance computing can be flexibly suitable for other different user environment.
Automating Web service composition is a challenging area of service oriented computing, requiring progress on a number of issues concerning the automation of control flow, the alignment of dataflow, verification, exe...
详细信息
Automating Web service composition is a challenging area of service oriented computing, requiring progress on a number of issues concerning the automation of control flow, the alignment of dataflow, verification, execution monitoring, etc. In this paper, we focus on aligning dataflow in semantic Web service composition to ensure the robustness when executing the composed service by preventing any cases when the wrong type of data is passed on from one service to the next. The proposed solution is unique in that it ensures the robustness of dataflow when automatically composing Web services through the use of agent-based negotiation between Web service providers. Starting from the semantic specification of a non-robust composition, we adapt the mechanism of concept abduction to compute the set of additional semantic descriptions necessary to ensure robustness of the semantic connections within the composed service. An agent-based negotiation is then used to solicit the additional semantic definitions which are required for achieving robust Web service compositions.
Red blood cells form rouleau at lower shear rates, convection being important in determining the rates of growth of these compliant and flexible structures. The rouleau can be disrupted into separated cells when the s...
详细信息
Red blood cells form rouleau at lower shear rates, convection being important in determining the rates of growth of these compliant and flexible structures. The rouleau can be disrupted into separated cells when the shear rates in flow are sufficiently high. Because relatively sophisticated computational models of cell-cell interactions in flow are now being advanced, with virtual-reality rendering for visualization, better quantitative information about rouleau formation is important for providing space-time images and aggregation-rate data of rouleau formation for comparison with such models. We provide here some such quantitative information on aggregation rate determined with a dilute RBC suspension observed in vitro in a flow chamber with a simple Hele-Shaw geometry. At a steady shear rate, following initial adhesion of some individual red cells to the glass wall, growth of wall-attached rouleau is observed while the number of single adherent red cells remains relatively constant. Rouleau adherent to the wall tend to align their major axis parallel to the steady flow.
Financial systems are nowadays dependent on electronic data processing technology. One of the most frequent causes of failure in this highly sensitive system are overvoltages mostly caused by lightning. This paper des...
详细信息
Financial systems are nowadays dependent on electronic data processing technology. One of the most frequent causes of failure in this highly sensitive system are overvoltages mostly caused by lightning. This paper describes how lightning electromagnetic pulse intrudes into computer information systems and causes system damage. Damage prevention is also discussed.
The n-dimensional grid is one of the most representative patterns of dataflow in parallel computation. The most frequently used scheduling models for grids is the unit execution-unit communication time (UET-UCT). We ...
详细信息
The n-dimensional grid is one of the most representative patterns of dataflow in parallel computation. The most frequently used scheduling models for grids is the unit execution-unit communication time (UET-UCT). We enhance the model of n-dimensional grid by adding extra diagonal edges. First, we calculate the optimal makespan for the generalized UET-UCT grid topology and then we establish the minimum number of processors required, to achieve the optimal makespan. Furthermore, we solve the scheduling problem for generalized n-dimensional grids by proposing an optimal time and space scheduling strategy. We thus prove that UET-UCT scheduling of generalized n-dimensional grids is low complexity tractable.
The accuracy of optical flow estimation algorithms has been improving steadily as evidenced by results on the Middlebury optical flow benchmark. The typical formulation, however, has changed little since the work of H...
详细信息
The accuracy of optical flow estimation algorithms has been improving steadily as evidenced by results on the Middlebury optical flow benchmark. The typical formulation, however, has changed little since the work of Horn and Schunck. We attempt to uncover what has made recent advances possible through a thorough analysis of how the objective function, the optimization method, and modern implementation practices influence accuracy. We discover that “classical” flow formulations perform surprisingly well when combined with modern optimization and implementation techniques. Moreover, we find that while median filtering of intermediate flow fields during optimization is a key to recent performance gains, it leads to higher energy solutions. To understand the principles behind this phenomenon, we derive a new objective that formalizes the median filtering heuristic. This objective includes a nonlocal term that robustly integrates flow estimates over large spatial neighborhoods. By modifying this new term to include information about flow and image boundaries we develop a method that ranks at the top of the Middlebury benchmark.
暂无评论