The Experts below are selected from a list of 2991 Experts worldwide ranked by ideXlab platform
Charles E Leiserson - One of the best experts on this subject based on the ideXlab platform.
-
can multithreaded programming save Massively Parallel Computing
International Parallel Processing Symposium, 1996Co-Authors: Charles E LeisersonAbstract:Massively Parallel Computing has taken a turn for the worse. MPP (Massively Parallel processor) companies have generally been doing poorly in the marketplace. The additional time to design and deliver MPP systems puts them a generation behind the latest small-scale microprocessor and SMP systems. Truly large machines have mean times to failure measured in days, limiting their ability to provide reliable Computing platforms for longrunning computations. Software for MPP’s is arcane, and porting a serial code from a conventional workstation to an MPP is a major chore, if not a research project. Is Massively Parallel Computing doomed? Does anybody care? We should care! Massively Parallel Computing is the only way to solve society’s most computationally intensive problems. In the last ten years, MPP’s have shown scientists and engineers from many disciplines that important problems they had previously considered beyond their reach are, in fact, solvable. The rapid advancement of electronics, automobile, and pharmaceutical designs demands ever-higher performance from simulation and analysis tools. The computational power needed for data mining and decision analysis is increasing at a rapid rate. The burgeoning popularity of the Internet is now making it possible to deliver high-performance Computing services to millions, if networks and software can meet the challenge. Algorithmic multithreaded programming, such as provided by the Cilk system being developed at MIT and the University of Texas at Austin, offers the hope of allowing Massively Parallel Computing to fulfill its promise, even if conventional MPP’s themselves fall by the technology wayside. Algorithmic multithreaded languages provide high-level Parallel abstractions for system resources-such as processors, memory, and files-thereby allowing the runtime system to map these abstract resources onto available physical resources dynamically, while providing solid guarantees of high performance. As a consequence, a program can execute adaptively and tolerate faults in a changeable Computing environment, such as the clusters of SMP workstations that appear to be the next high-performance fad. Moreover, a multithreaded program can “scale down” to run on a single processor with the same performance as serial C code, thereby removing a major barrier between Parallel and serial programming. Significant problems remain before multithreading can replace the existing base of Parallel software, however. The most pressing appears to be the problem of duplicating the successes of data Parallelism and message passing for problems that require tight and frequent synchronization. In addition, multithreading will demand stronger support from architectures and operating systems for low-latency interrupts and low-latency inter-processor communication. Proceedings of the 10th International Parallel Processing Symposium (IPPS '96) 1063-7133/96 $10.00 © 1996 IEEE
-
IPPS - Can Multithreaded Programming Save Massively Parallel Computing
1996Co-Authors: Charles E LeisersonAbstract:Massively Parallel Computing has taken a turn for the worse. MPP (Massively Parallel processor) companies have generally been doing poorly in the marketplace. The additional time to design and deliver MPP systems puts them a generation behind the latest small-scale microprocessor and SMP systems. Truly large machines have mean times to failure measured in days, limiting their ability to provide reliable Computing platforms for longrunning computations. Software for MPP’s is arcane, and porting a serial code from a conventional workstation to an MPP is a major chore, if not a research project. Is Massively Parallel Computing doomed? Does anybody care? We should care! Massively Parallel Computing is the only way to solve society’s most computationally intensive problems. In the last ten years, MPP’s have shown scientists and engineers from many disciplines that important problems they had previously considered beyond their reach are, in fact, solvable. The rapid advancement of electronics, automobile, and pharmaceutical designs demands ever-higher performance from simulation and analysis tools. The computational power needed for data mining and decision analysis is increasing at a rapid rate. The burgeoning popularity of the Internet is now making it possible to deliver high-performance Computing services to millions, if networks and software can meet the challenge. Algorithmic multithreaded programming, such as provided by the Cilk system being developed at MIT and the University of Texas at Austin, offers the hope of allowing Massively Parallel Computing to fulfill its promise, even if conventional MPP’s themselves fall by the technology wayside. Algorithmic multithreaded languages provide high-level Parallel abstractions for system resources-such as processors, memory, and files-thereby allowing the runtime system to map these abstract resources onto available physical resources dynamically, while providing solid guarantees of high performance. As a consequence, a program can execute adaptively and tolerate faults in a changeable Computing environment, such as the clusters of SMP workstations that appear to be the next high-performance fad. Moreover, a multithreaded program can “scale down” to run on a single processor with the same performance as serial C code, thereby removing a major barrier between Parallel and serial programming. Significant problems remain before multithreading can replace the existing base of Parallel software, however. The most pressing appears to be the problem of duplicating the successes of data Parallelism and message passing for problems that require tight and frequent synchronization. In addition, multithreading will demand stronger support from architectures and operating systems for low-latency interrupts and low-latency inter-processor communication. Proceedings of the 10th International Parallel Processing Symposium (IPPS '96) 1063-7133/96 $10.00 © 1996 IEEE
John Geweke - One of the best experts on this subject based on the ideXlab platform.
-
adaptive sequential posterior simulators for Massively Parallel Computing environments
arXiv: Computation, 2013Co-Authors: Garland Durham, John GewekeAbstract:Massively Parallel desktop Computing capabilities now well within the reach of individual academics modify the environment for posterior simulation in fundamental and potentially quite advantageous ways. But to fully exploit these benefits algorithms that conform to Parallel Computing environments are needed. Sequential Monte Carlo comes very close to this ideal whereas other approaches like Markov chain Monte Carlo do not. This paper presents a sequential posterior simulator well suited to this Computing environment. The simulator makes fewer analytical and programming demands on investigators, and is faster, more reliable and more complete than conventional posterior simulators. The paper extends existing sequential Monte Carlo methods and theory to provide a thorough and practical foundation for sequential posterior simulation that is well suited to Massively Parallel Computing environments. It provides detailed recommendations on implementation, yielding an algorithm that requires only code for simulation from the prior and evaluation of prior and data densities and works well in a variety of applications representative of serious empirical work in economics and finance. The algorithm is robust to pathological posterior distributions, generates accurate marginal likelihood approximations, and provides estimates of numerical standard error and relative numerical efficiency intrinsically. The paper concludes with an application that illustrates the potential of these simulators for applied Bayesian inference.
-
adaptive sequential posterior simulators for Massively Parallel Computing environments
Advances in Econometrics, 2013Co-Authors: Garland Durham, John GewekeAbstract:Massively Parallel desktop Computing capabilities now well within the reach of individual academics modify the environment for posterior simulation in fundamental and potentially quite advantageous ways. But to fully exploit these benfits algorithms that conform to Parallel Computing environments are needed. Sequential Monte Carlo comes very close to this ideal whereas other approaches like Markov chain Monte Carlo do not. This paper presents a sequential posterior simulator well suited to this Computing environment. The simulator makes fewer analytical and programming demands on investigators, and is faster, more reliable and more complete than conventional posterior simulators. The paper extends existing sequential Monte Carlo methods and theory to provide a thorough and practical foundation for sequential posterior simulation that is well suited to Massively Parallel Computing environments. It provides detailed recommendations on implementation, yielding an algorithm that requires only code for simulation from the prior and evaluation of prior and data densities and works well in a variety of applications representative of serious empirical work in economics and finance. The algorithm is robust to pathological posterior distributions, generates accurate marginal likelihood approximations, and provides estimates of numerical standard error and relative numerical efficiency intrinsically. The paper concludes with an application that illustrates the potential of these simulators for applied Bayesian inference.
Thanavat Junchaya - One of the best experts on this subject based on the ideXlab platform.
-
An integrated route assignment and traffic simulation system with Massively Parallel Computing architecture
1995Co-Authors: Thanavat JunchayaAbstract:This research presents two critical issues in the development of an integrated route assignment and traffic simulation system of ATMS-ATIS applications. The first issue addresses the conceptual and algorithmic aspects of the models. the conceptual aspects of the models. The second concerns with computation and implementation efficiency which lead to the exploration of using advanced Parallel Computing architecture. We propose an integrated system that has been implemented on a Massively Parallel Computing architecture. This paper presents the structure of the proposed system, along with a brief description of each component.
-
a user optimum route navigation model with a Massively Parallel Computing architecture
Transportation Planning and Technology, 1994Co-Authors: Ganglen Chang, Thanavat Junchaya, Leimin ZhuangAbstract:This paper presents a user‐optimum route assignment model which has been developed primarily for ATMS/ ATIS real‐time applications. The model uses a learning process in which the new route assignment uses information gained from previous iterations in the computation of new paths. During each iteration, the vehicle equipped with in‐vehicle navigation systems are assigned routes according to time‐dependent shortest paths. These assigned paths will be used to predict future traffic flow patterns and link trip times. Differing from existing assignment‐simulation model frameworks which require an external simulation model to estimate link travel time, the proposed model consists of a simulation‐like loading mechanism to estimate link trip time so as to take into account signal control and queues. Furthermore, the route assignment model for a Massively Parallel computer has been developed for real‐time applications. Several numerical examples have been carried out to examine the properties of the proposed mode...
-
simulating network traffic flows with a Massively Parallel Computing architecture
Winter Simulation Conference, 1993Co-Authors: Ganglen Chang, Thanavat JunchayaAbstract:This paper presents a traffic network simulation model for real-time applications in IVHS. The proposed model has incorporated three key features essential for IVHS operations: (1) the capability of simulating both freeways and surface street networks as an integrated network; (2) a path-processing capability for representing drivers' route choice behavior at an individual/vehicle level; and (3) the capability of simulating different subnetworks at different levels of detail so as to increase the execution speed for real-time operations. Three simulation methodologies are implemented in the real-time traffic simulation model to satisfy the above requirements. These three methodologies are macroparticle traffic simulation model (MPSM), modified NIPSM (M-MPSM), and microscopic (MICRO) model. Due to the real-time operating requirement, the proposed simulation model has been Parallelized to take advantage of the Parallel computers since they can offer the required computational power at an economical cost/performance ratio. Several simulation experiments have been carried out to compare the execution speed of each methodology. The preliminary research results indicate that Parallel Computing architecture offers a very promising alternative for the implementation of real-time traffic simulation.
-
PACIFIC RIM TRANSTECH CONFERENCE PROCEEDINGS. VOLUME I. AN INTEGRATED ROUTE ASSIGNMENT AND TRAFFIC SIMULATION SYSTEM WITH A Massively Parallel Computing ARCHITECTURE
1993Co-Authors: G-l Chang, Thanavat Junchaya, L ZhuangAbstract:Two critcal issues in the development of an integrated route assignment and traffic simulation system for ATMS (advanced traffic management systems) - ATIS (advanced traveler information systems) applications are researched. The first case addresses the conceptual and algorithmic aspects of the models. The second isse is concerned with computation and implementation efficiency and the exploration of advanced Parallel Computing architecture. An integrated system is proposed that is implemented on a Massively Parallel Computing architecture.
-
exploring real time traffic simulation with Massively Parallel Computing architecture
Transportation Research Part C-emerging Technologies, 1993Co-Authors: Thanavat Junchaya, Ganglen ChangAbstract:Abstract The advent of Parallel Computing architectures presents an attainable opportunity for transportation professionals to simulate a large-scale traffic network with sufficiently fast response time for real-time operation. However, it necessitates a fundamental change in the modelling algorithm to take full advantage of Parallel Computing. Currently there are two general types of Parallel processing architectures: (a) single instruction multiple data (SIMD) streams, and (b) multiple instruction multiple data streams (MIMD). This paper describes a model to simulate network traffic with the Connection Machine, a Massively Parallel SIMD computer. First we introduce the basic Parallel Computing architectures along with a list of commercially available Parallel computers. It is followed by an in-depth presentation of the proposed simulation methodology with a Massively Parallel computer. The proposed traffic simulation model has an inherent path-processing capability to represent drivers' route choice behavior at the individual/vehicle level. Such a feature is critical to its integration with a real-time dynamic assignment model in IVHS applications. The proposed model has been implemented on the Connection Machine. Several simulation experiments were carried out which show that Massively Parallel computers provide a viable alternative for use in the real-time application. The results show that the CM-2 with 16,384 processors can simulate 32,000 vehicles for 30 minutes at a one-second interval within 3 1 2 minutes.
Ganglen Chang - One of the best experts on this subject based on the ideXlab platform.
-
a user optimum route navigation model with a Massively Parallel Computing architecture
Transportation Planning and Technology, 1994Co-Authors: Ganglen Chang, Thanavat Junchaya, Leimin ZhuangAbstract:This paper presents a user‐optimum route assignment model which has been developed primarily for ATMS/ ATIS real‐time applications. The model uses a learning process in which the new route assignment uses information gained from previous iterations in the computation of new paths. During each iteration, the vehicle equipped with in‐vehicle navigation systems are assigned routes according to time‐dependent shortest paths. These assigned paths will be used to predict future traffic flow patterns and link trip times. Differing from existing assignment‐simulation model frameworks which require an external simulation model to estimate link travel time, the proposed model consists of a simulation‐like loading mechanism to estimate link trip time so as to take into account signal control and queues. Furthermore, the route assignment model for a Massively Parallel computer has been developed for real‐time applications. Several numerical examples have been carried out to examine the properties of the proposed mode...
-
simulating network traffic flows with a Massively Parallel Computing architecture
Winter Simulation Conference, 1993Co-Authors: Ganglen Chang, Thanavat JunchayaAbstract:This paper presents a traffic network simulation model for real-time applications in IVHS. The proposed model has incorporated three key features essential for IVHS operations: (1) the capability of simulating both freeways and surface street networks as an integrated network; (2) a path-processing capability for representing drivers' route choice behavior at an individual/vehicle level; and (3) the capability of simulating different subnetworks at different levels of detail so as to increase the execution speed for real-time operations. Three simulation methodologies are implemented in the real-time traffic simulation model to satisfy the above requirements. These three methodologies are macroparticle traffic simulation model (MPSM), modified NIPSM (M-MPSM), and microscopic (MICRO) model. Due to the real-time operating requirement, the proposed simulation model has been Parallelized to take advantage of the Parallel computers since they can offer the required computational power at an economical cost/performance ratio. Several simulation experiments have been carried out to compare the execution speed of each methodology. The preliminary research results indicate that Parallel Computing architecture offers a very promising alternative for the implementation of real-time traffic simulation.
-
exploring real time traffic simulation with Massively Parallel Computing architecture
Transportation Research Part C-emerging Technologies, 1993Co-Authors: Thanavat Junchaya, Ganglen ChangAbstract:Abstract The advent of Parallel Computing architectures presents an attainable opportunity for transportation professionals to simulate a large-scale traffic network with sufficiently fast response time for real-time operation. However, it necessitates a fundamental change in the modelling algorithm to take full advantage of Parallel Computing. Currently there are two general types of Parallel processing architectures: (a) single instruction multiple data (SIMD) streams, and (b) multiple instruction multiple data streams (MIMD). This paper describes a model to simulate network traffic with the Connection Machine, a Massively Parallel SIMD computer. First we introduce the basic Parallel Computing architectures along with a list of commercially available Parallel computers. It is followed by an in-depth presentation of the proposed simulation methodology with a Massively Parallel computer. The proposed traffic simulation model has an inherent path-processing capability to represent drivers' route choice behavior at the individual/vehicle level. Such a feature is critical to its integration with a real-time dynamic assignment model in IVHS applications. The proposed model has been implemented on the Connection Machine. Several simulation experiments were carried out which show that Massively Parallel computers provide a viable alternative for use in the real-time application. The results show that the CM-2 with 16,384 processors can simulate 32,000 vehicles for 30 minutes at a one-second interval within 3 1 2 minutes.
Garland Durham - One of the best experts on this subject based on the ideXlab platform.
-
adaptive sequential posterior simulators for Massively Parallel Computing environments
arXiv: Computation, 2013Co-Authors: Garland Durham, John GewekeAbstract:Massively Parallel desktop Computing capabilities now well within the reach of individual academics modify the environment for posterior simulation in fundamental and potentially quite advantageous ways. But to fully exploit these benefits algorithms that conform to Parallel Computing environments are needed. Sequential Monte Carlo comes very close to this ideal whereas other approaches like Markov chain Monte Carlo do not. This paper presents a sequential posterior simulator well suited to this Computing environment. The simulator makes fewer analytical and programming demands on investigators, and is faster, more reliable and more complete than conventional posterior simulators. The paper extends existing sequential Monte Carlo methods and theory to provide a thorough and practical foundation for sequential posterior simulation that is well suited to Massively Parallel Computing environments. It provides detailed recommendations on implementation, yielding an algorithm that requires only code for simulation from the prior and evaluation of prior and data densities and works well in a variety of applications representative of serious empirical work in economics and finance. The algorithm is robust to pathological posterior distributions, generates accurate marginal likelihood approximations, and provides estimates of numerical standard error and relative numerical efficiency intrinsically. The paper concludes with an application that illustrates the potential of these simulators for applied Bayesian inference.
-
adaptive sequential posterior simulators for Massively Parallel Computing environments
Advances in Econometrics, 2013Co-Authors: Garland Durham, John GewekeAbstract:Massively Parallel desktop Computing capabilities now well within the reach of individual academics modify the environment for posterior simulation in fundamental and potentially quite advantageous ways. But to fully exploit these benfits algorithms that conform to Parallel Computing environments are needed. Sequential Monte Carlo comes very close to this ideal whereas other approaches like Markov chain Monte Carlo do not. This paper presents a sequential posterior simulator well suited to this Computing environment. The simulator makes fewer analytical and programming demands on investigators, and is faster, more reliable and more complete than conventional posterior simulators. The paper extends existing sequential Monte Carlo methods and theory to provide a thorough and practical foundation for sequential posterior simulation that is well suited to Massively Parallel Computing environments. It provides detailed recommendations on implementation, yielding an algorithm that requires only code for simulation from the prior and evaluation of prior and data densities and works well in a variety of applications representative of serious empirical work in economics and finance. The algorithm is robust to pathological posterior distributions, generates accurate marginal likelihood approximations, and provides estimates of numerical standard error and relative numerical efficiency intrinsically. The paper concludes with an application that illustrates the potential of these simulators for applied Bayesian inference.