Filters








447 Hits in 2.1 sec

Automatic node selection for high performance applications on networks

Jaspal Subhlok, Peter Lieu, Bruce Lowekamp
1999 Proceedings of the seventh ACM SIGPLAN symposium on Principles and practice of parallel programming - PPoPP '99  
Under the network conditions used for experiments, the increase in execution time due to compute loads and network congestion was reduced by half with node selection.  ...  A central problem in executing performance critical parallel and distributed applications on shared networks is the selection of computation nodes and communication paths for execution.  ...  Remos API Remos is an API designed to support the development of network-aware parallel and distributed applications on diverse network architectures.  ... 
doi:10.1145/301104.301119 dblp:conf/ppopp/SubhlokLL99 fatcat:qhevhcx65va2tdokhn7sgcfmgm

Automatic node selection for high performance applications on networks

Jaspal Subhlok, Peter Lieu, Bruce Lowekamp
1999 SIGPLAN notices  
Under the network conditions used for experiments, the increase in execution time due to compute loads and network congestion was reduced by half with node selection.  ...  A central problem in executing performance critical parallel and distributed applications on shared networks is the selection of computation nodes and communication paths for execution.  ...  Remos API Remos is an API designed to support the development of network-aware parallel and distributed applications on diverse network architectures.  ... 
doi:10.1145/329366.301119 fatcat:3icvyz6zfngxli6ny2baj2ncpi

Network performance-aware collective communication for clustered wide-area systems

Thilo Kielmann, Henri E. Bal, Sergei Gorlatch, Kees Verstoep, Rutger F.H. Hofman
2001 Parallel Computing  
network performance-aware) Outline of the talk: • The performance model P-LogP transit time for single bit os send overhead time, sender is busy with msg or receive overhead time, receiver is busy with  ...  completion time for clustered systems (making MagPIe network performance-aware) Goal of this work Designing MPI's collective operations with minimal completion time for clustered systems (making MagPIe  ... 
doi:10.1016/s0167-8191(01)00098-9 fatcat:n2zklug6ifc7xo47ajroj7i6vy

Resource-Aware Application State Monitoring

Shicong Meng, Srinivas Raghav Kashyap, Chitra Venkatramani, Ling Liu
2012 IEEE Transactions on Parallel and Distributed Systems  
In this paper, we present REMO, a REsource-aware application state MOnitoring system, to address the challenge of monitoring overlay construction.  ...  Moreover, we enhance REMO in terms of both performance and applicability with a series of optimization and extension techniques.  ...  ACKNOWLEDGMENTS This work is partially supported by NSF grants from CISE CyberTrust program, NetSE program, and CyberTrust Cross Cutting program, and an IBM faculty award and an Intel ISTC grant on Cloud Computing  ... 
doi:10.1109/tpds.2012.82 fatcat:xy7b35mqp5fnjg24hta6goyz5y

Combining active and passive network measurements to build scalable monitoring systems on the grid

Bruce B. Lowekamp
2003 Performance Evaluation Review  
Because the network provides the wires that connect a grid, understanding the performance provided by a network is crucial to achieving satisfactory performance from many grid applications.  ...  This paper is a preliminary report on the Wren project, which is focused on developing scalable solutions for network performance monitoring.  ...  Additionally, the computation itself may be parallel and relocatable, with only the source and destination fixed, but the network and computational path used along the way open to adaptation.  ... 
doi:10.1145/773056.773061 fatcat:avipmmp4yvafzh2y46lp4og2vu

Deep Learning for Climate Model Output Statistics [article]

Michael Steininger, Daniel Abel, Katrin Ziegler, Anna Krause, Heiko Paeth, Andreas Hotho
2020 arXiv   pre-print
In this work, we explore the feasibility and potential of deep learning with convolutional neural networks (CNNs) for MOS.  ...  We propose the CNN architecture ConvMOS specifically designed for reducing errors in climate model outputs and apply it to the climate model REMO.  ...  Despite its inherent randomness, the RF method is only calculated once since this computation already took over four days for our study area with 15 CPU cores in parallel.  ... 
arXiv:2012.10394v1 fatcat:jlfhs5bmkrexjmnfkpq3no7zra

Mesh Generation and Optimistic Computation on the Grid [chapter]

Nikos Chrisochoides, Craig Lee, Bruce Lowekamp
2004 Performance Analysis and Grid Computing  
Mesh generation is the primary test application where boundary mesh cavities can be optimistically expanded in parallel.  ...  This paper describes the concept of optimistic grid computing.  ...  Optimistic Computation for Grid Environments General Optimistic Computation Optimistic parallel computation is very similar to optimistic parallel simulation [19] .  ... 
doi:10.1007/978-1-4615-0361-3_14 fatcat:pvlbqxkzubewxnewt2azufvn3m

Adaptive Communication Algorithms for Distributed Heterogeneous Systems

Prashanth B. Bhat, Viktor K. Prasanna, C.S. Raghavendra
1999 Journal of Parallel and Distributed Computing  
Efficient algorithms for various collective communication patterns have been developed for tightly coupled parallel architectures with homogeneous networks [2, 19] .  ...  Further, these are static algorithms, with no provision for adaptivity to network conditions. In Section 3, we introduce our approach for developing network-aware communication techniques.  ...  Greedy Technique The greedy technique is an approximation to the matching technique, with a lower computational complexity.  ... 
doi:10.1006/jpdc.1999.1571 fatcat:gj75ddzwbvc27ka4lvyofybosa

Identifying Logical Homogeneous Clusters for Efficient Wide-Area Communications [chapter]

Luiz Angelo Barchet-Estefanel, Grégory Mounié
2004 Lecture Notes in Computer Science  
Recently, many works focus on the implementation of collective communication operations adapted to wide area computational systems, like computational Grids or global-computing.  ...  In this paper, we describe a strategy to gather network information from different local-area networks and to construct "logical homogeneous clusters", better suited to the performance modelling.  ...  It is also a true concern for users that run parallel applications over their LANs, where there can be combined different machines and network supports.  ... 
doi:10.1007/978-3-540-30218-6_45 fatcat:rlixkglkorg4lpl5f7s22k5rfi

Trajectory Pattern Mining [chapter]

Hoyoung Jeung, Man Lung Yiu, Christian S. Jensen
2011 Computing with Spatial Trajectories  
In step with the rapidly growing volumes of available moving-object trajectory data, there is also an increasing need for techniques that enable the analysis of trajectories.  ...  As an example, Figure 5 .4(a) illustrates a road network, with its graph model shown in Figure 5 .4(b).  ...  The resulting location-aware devices find widespread use in various business and personal settings in society.  ... 
doi:10.1007/978-1-4614-1629-6_5 fatcat:ftpmjhbabbewpii6parc6kklji

GridRM: A Resource Monitoring Architecture for the Grid [chapter]

Mark Baker, Garry Smith
2002 Lecture Notes in Computer Science  
The resources of interest can include all manner of networked devices, from a remote sensor or satellite feed through to a computational node or a communications link.  ...  Remos The Remos (REsource MOnitoring System) [8] allows network-aware applications to obtain information about their distributed execution environment by providing an interface that hides system heterogeneity  ...  Autopilot The Autopilot project, based on the Pablo Toolkit [7], is an infrastructure for real-time adaptive control of parallel and distributed computing resources.  ... 
doi:10.1007/3-540-36133-2_24 fatcat:qe5ksymmb5fnladic754h3cxfu

Analyzing Relative Motion within Groups ofTrackable Moving Point Objects [chapter]

Patrick Laube, Stephan Imfeld
2002 Lecture Notes in Computer Science  
The analysis concept called REMO (RElative MOtion) is based on the comparison of motion parameters of objects over time.  ...  Whereas recent efforts of dealing with dynamic phenomena within the GIScience community mainly concentrated on modeling and representation, this research project concentrates on the analytic task.  ...  This measure is computed for every object every time step and accumulated over the objects.  ... 
doi:10.1007/3-540-45799-2_10 fatcat:xwkqnfd2mrawbaqkvwtppntz64

Assessing the Quality of Automatically Built Network Representations

Lionel Eyraud-Dubois, Martin Quinson
2007 Seventh IEEE International Symposium on Cluster Computing and the Grid (CCGrid '07)  
In order to efficiently use Grid resources, users or middlewares must use some network information, and in particular some knowledge of the platform network.  ...  As such knowledge is usually not available, one must use tools which automatically build a topological network model through some measurements.  ...  Qualitative information such as the network topology is crucial to achieve such tasks as running network-aware applications [2] , efficiently placing servers [3] , or predicting and optimizing collective  ... 
doi:10.1109/ccgrid.2007.25 dblp:conf/ccgrid/Eyraud-DuboisQ07 fatcat:rnjr7rjczjepffpljpoqeqxvba

A taxonomy of grid monitoring systems

Serafeim Zanikolas, Rizos Sakellariou
2005 Future generations computer systems  
The paper concludes with, among others, a discussion of the considered systems, as well as directions for future research.  ...  Remos (L2b.N.S) The Resource Monitoring System (Remos) provides to network-aware applications an application programming interface (API) for run-time enquiry of performance measurements of local and wide  ...  In contrast, a network topology graph provides a logical view of the physical interconnection between compute and network nodes, annotated with information on link capacity, current bandwidth utilisation  ... 
doi:10.1016/j.future.2004.07.002 fatcat:flzstw3rvvhcvh5kot7odlnmsa

Characterizing NAS benchmark performance on shared heterogeneous networks

J. Subhlok, S. Venkataramaiah, A. Singh
2002 Proceedings 16th International Parallel and Distributed Processing Symposium  
The goal of this research is to develop performance profiles of parallel and distributed applications in order to predict their execution time under different network conditions.  ...  The results are used to characterize the behavior of the NAS benchmark programs with resource sharing.  ...  Support was also provided by the Texas Advanced Technology Program under grant number 003652-0424 and the University of Houston's Texas Learning and Computation Center.  ... 
doi:10.1109/ipdps.2002.1015659 dblp:conf/ipps/SubhlokVS02 fatcat:7vugl6zoz5el7grfl4mozd4liu
« Previous Showing results 1 — 15 out of 447 results