In this paper, we analyze the performance of the prototype software developed for the ATLAS Second-Level Trigger. The 00 framework written in C++ has been used to implement a distributed system which collects (simulat...
详细信息
In this paper, we analyze the performance of the prototype software developed for the ATLAS Second-Level Trigger. The 00 framework written in C++ has been used to implement a distributed system which collects (simulated) detector data on which it executes event selection algorithms. The software has been used on testbeds of up to 100 nodes with various interconnect technologies. The final system will have to sustain traffic of similar to40 Gb/s and require an estimated number of similar to750 processors. Timing measurements are crucial for issues such as trigger decision latency, assessment of required CPU and network capacity, scalability, and load-balancing. In addition, final architectural and technological choices, code optimization, and system tuning require a detailed understanding of both CPU utilization and trigger decision latency. In this paper, we describe the instrumentation used to disentangle effects due to such factors as OS system intervention, blocking on interlocks (applications are multithreaded), multiple CPUs, and I/O. This is followed by an analysis of the measurements and concluding with suggestions for improvements to the ATLAS T)rigger/DAQ dataflow components in the next phase of the project.
Future generation of microprocessors will require highperformance Voltage Regulator Modules (VRMs) that produce tightly regulated low supply voltage with very small deviation window and able to respond very quickly t...
详细信息
ISBN:
(纸本)0780374029
Future generation of microprocessors will require highperformance Voltage Regulator Modules (VRMs) that produce tightly regulated low supply voltage with very small deviation window and able to respond very quickly to a large and continuous load transients at high output current while maintaining a high power density. On one hand, the current drawn from the VRM by the microprocessor is continuously changing since it depends at the current use of the microprocessor. On the other hand, high Frequency VRMs as any Power Electronics System is a complex combination of linear, nonlinear, and switching elements that is required to have fast dynamics. Moreover, this complex combination is also real-time system that needs to continuously and instantly monitor and respond to the load changes (the microprocessor). A highperformance basic control loop is essential to follow up with such transients. Such controller design is usually complicated especially since it requires high knowledge of the converter and its behavior and accurate converter model that includes nonlinearities and parameters and components variations. DSP has many advantages over the analog circuits when it comes to applying highperformance sophisticated control techniques such the simplicity in applying sophisticated control algorithms and modifying them via software revision, lower environmental and noise sensitivity, and less components count.
In the pastfew years, increasing interest has been shown in using Java as a language for performance-oriented distributed and parallel computing. Most Java-based systems that support portable parallel and distributed ...
详细信息
ISBN:
(纸本)0769517455
In the pastfew years, increasing interest has been shown in using Java as a language for performance-oriented distributed and parallel computing. Most Java-based systems that support portable parallel and distributed computing either require the programmer to deal with intricate low-level details of Java which can be a tedious, time-consuming and error-prone task, or prevent the programmer from controlling locality of data. In contrast to most existing systems, JavaSymphony - a class library written entirely in Java - allows to control parallelism, load balancing, and locality, at a high level. Objects can be explicitly distributed and migrated based on virtual architectures which impose a virtual hierarchy on a distributed/parallel system of physical computing nodes. The concept of blocking/nonblocking remote method invocation is used to exchange data among distributed objects and to process work by remote objects. In this paper we evaluate the JavaSymphony programming API for a variety, of distributed/parallel algorithms which comprises backtracking, N-body, encryption/decryption algorithms and asynchronous nested optimizationalgorithms. performance results are presented for both homogeneous and heterogeneous cluster architectures. Moreover we compare JavaSymphony with an alternative well-known semi-automatic system.
software pipelining is widely used as a compiler optimization technique to achieve highperformance in machines that exploit instruction-level parallelism. However, surprisingly, there have been few theoretical or emp...
详细信息
software pipelining is widely used as a compiler optimization technique to achieve highperformance in machines that exploit instruction-level parallelism. However, surprisingly, there have been few theoretical or empirical results on optimal software pipelining of loops with control flows. In this paper, we present three new contributions for this under-investigated problem. First, we propose a necessary and sufficient condition for a loop with control flows to have an optimally software-pipelined program. We also present a decision procedure to compute the condition. Second, we present two software pipelining algorithms. The first algorithm computes an optimal solution for every loop satisfying the condition, but may run in exponential time. The second algorithm computes optimal solutions efficiently for most (but not all) loops satisfying the condition. Third, we present experimental results which strongly indicate that achieving the optimality in the software-pipelined programs is a viable goal in practice with realistic hardware support.
Optimal gait planning is applied in this work to the problem of improving stability in quadruped locomotion. In many settings, it is desired to operate legged machines at highperformance levels where rapid velocities...
详细信息
Optimal gait planning is applied in this work to the problem of improving stability in quadruped locomotion. In many settings, it is desired to operate legged machines at highperformance levels where rapid velocities and a changing environment make stability of utmost concern. Since gait planning still remains a vital component of legged system control design, an efficient method of determining periodic paths is presented which optimize a dynamic stability criterion. Efficient recursive multibody algorithms are used with numerical optimal control software to solve the minimax performance stability criteria.
The proceedings contain 128 papers. The special focus in this conference is on Computational Geometry and Applications;Computing in Medicine;highperformance Computing in Particle Accelerator Science and Technology. T...
ISBN:
(纸本)3540435948
The proceedings contain 128 papers. The special focus in this conference is on Computational Geometry and Applications;Computing in Medicine;highperformance Computing in Particle Accelerator Science and Technology. The topics include: Recent Developments in Motion Planning;Extreme Distances in Multicolored Point Sets;Balanced Partition of Minimum Spanning Trees;On the Quality of Partitions Based on Space-Filling Curves;The Largest Empty Annulus Problem;Mapping Graphs on the Sphere to the Finite Plane;Improved Optimal Weighted Links algorithms;A Linear Time Heuristics for Trapezoidation of GIS Polygons;The Morphology of Building Structures;Voronoi and Radical Tessellations of Packings of Spheres;Collision Detection optimization in a Multi-particle System;optimization Techniques in an Event-Driven Simulation of a Shaker Ball Mill;Modified DAG Location for Delaunay Triangulation;TIN Meets CAD;Extracting Meaningful Slopes from Terrain Contours;Duality in Disk Induced Flows;Improvement of Digital Terrain Model Interpolation Using SFS Techniques with Single Satellite Imagery;Implementing an Augmented Scene Delivery System;Inspection Strategies for Complex Curved Surfaces Using CMM;The Free Form Deformation of Phytoplankton Models;Curvature Based Registration with Applications to MR-Mammography;Full Scale nonlinear Electromagnetic Inversion for Biological Objects;Propagation of Excitation Waves and Their Mutual Interactions in the Surface Layer of the Ball with Fast Accessory Paths and the Pacemaker;Computing Optimal Trajectories for Medical Treatment Planning and optimization;CAD Recognition Using Three Mathematical Models and 3D Quantification Visualization of Vascular Structures in Magnetic Resonance Angiographic Images.
Presents the algorithms to simulate nonlinear kinematics of a 5-axis milling machine. The simulator is based on 3D representation and employing the inverse kinematics approach to derive the corresponding rotational an...
详细信息
Presents the algorithms to simulate nonlinear kinematics of a 5-axis milling machine. The simulator is based on 3D representation and employing the inverse kinematics approach to derive the corresponding rotational and translation movement of the mechanism. The simulator makes it possible to analyze the accuracy of a 3D tool-path based on a prescribed set of the cutter location (CL) points as well as a set of the cutter contact (CC) points with tool inclination angle. The resulting trajectory of the tool is not unique and depends on the initial set up of the machine which in turn is problem dependent. Furthermore, the simulator can be used to simulate the milling process, verify the final cut and estimate the errors of the actual tool-path before the real workpiece is actually being tested with the real machine. Thus, reducing the cost of iterative trial and error. Tool path simulation is verified by a series of cutting experiments performed by means of the proposed software and evaluates the accuracy of milling. It has been shown that the proposed graphical 3D software presents an efficient interactive approach to the interactive modification of a tool path based on an appropriate set of transformations as well as to verification of the tool path optimizationalgorithms.
Widespread acceptance of Condition Based Monitoring (CBM) systems has been hampered by, among other things, high costs and inaccurate diagnostics. The advent of new methods for signal processing, local wireless networ...
详细信息
ISBN:
(纸本)0819444502
Widespread acceptance of Condition Based Monitoring (CBM) systems has been hampered by, among other things, high costs and inaccurate diagnostics. The advent of new methods for signal processing, local wireless networks, and an industry standard architecture is an opportunity to develop low cost, reliable, practical health monitoring systems. We will discuss the signal processing issues that contribute to poor performance and how new algorithms can provide near optimum detection and recognition of broad-band signals in nonideal, changing, noise. We will then discuss the ONR sponsored Open Systems Architecture (OSA) and how that provides a common operating protocol for health monitoring systems. The heart of the OSA/CBM system is protocols for communication among the hardware and software components of a generalized CBM system which allows rapid and easy integration of specialty components. Finally, we will discuss the impact of several new technologies, including local wireless networks. For example, in many potential installations, almost 90% of the cost of installation is the wiring from the sensors to the processing units. By processing the raw data at the sensor and using a local wireless network to move data and monitor the CBM system itself, the cost of health monitoring can be dramatically reduced.
Photoresist (PR) outgassing during implantation can be significant for high current, high energy implants, and if ignored, can result in dose variations. The SWIFT ion implanter uses an intelligent software algorithm ...
详细信息
Photoresist (PR) outgassing during implantation can be significant for high current, high energy implants, and if ignored, can result in dose variations. The SWIFT ion implanter uses an intelligent software algorithm in conjunction with fast beam sampling to allow the system to reliably compensate for photoresist outgassing, allowing excellent wafer-to-wafer dose repeatability. In addition, the beam charge recovery tolerance has been made configurable, to allow easy optimization of system throughput versus dose control. In this paper, performance data is presented for photoresist wafers implanted using the intelligent software algorithm compared to bare wafers implanted under identical conditions.
Future generation of microprocessors will require highperformance Voltage Regulator Modules (VRMs) that produce tightly regulated low supply voltage with very small deviation window and able to respond very quickly t...
详细信息
Future generation of microprocessors will require highperformance Voltage Regulator Modules (VRMs) that produce tightly regulated low supply voltage with very small deviation window and able to respond very quickly to a large and continuous load transients at high output current while maintaining a high power density. On one hand, the current drawn from the VRM by the microprocessor is continuously changing since it depends at the current use of the microprocessor. On the other hand, high Frequency VRMs as any Power Electronics System is a complex combination of linear, nonlinear, and switching elements that is required to have fast dynamics. Moreover, this complex combination is also real-time system that needs to continuously and instantly monitor and respond to the load changes (the microprocessor). A highperformance basic control loop is essential to follow up with such transients. Such controller design is usually complicated especially since it requires high knowledge of the converter and its behavior and accurate converter model that includes nonlinearities and parameters and components variations. DSP has many advantages over the analog circuits when it comes to applying highperformance sophisticated control techniques such the simplicity in applying sophisticated control algorithms and modifying them via software revision, lower environmental and noise sensitivity, and less components count. In this paper, a DSP setup to be
暂无评论