Filters








989 Hits in 6.1 sec

Enabling efficient and flexible coupling of parallel scientific applications

Li Zhang, M. Parashar
2006 Proceedings 20th IEEE International Parallel & Distributed Processing Symposium  
It enables efficient computation of communication schedules, supports low-overheads processor-to-processor data streaming, and provides high-level abstraction for application developers.  ...  Emerging scientific and engineering simulations are presenting challenging requirements for coupling between multiple physics models and associated parallel codes that execute independently and in a distributed  ...  An alternate approach embeds the parallel data redistribution support into a Parallel Remote Method Invocation (PRMI) [10] mechanism.  ... 
doi:10.1109/ipdps.2006.1639354 dblp:conf/ipps/ZhangP06 fatcat:kwvct2raj5eapeeyxwkll3uigy

Elastic Remote Methods [chapter]

K. R. Jayaram
2013 Lecture Notes in Computer Science  
method invocations with respect to the application's workload, without requiring major changes to clients (invokers) of remote methods, (2) enables flexible elastic scaling by allowing developers to use  ...  centers/IaaS clouds through Apache Mesos.  ...  Having shared state and mutual exclusion through locks or synchronized methods further decreases parallelism.  ... 
doi:10.1007/978-3-642-45065-5_8 fatcat:zi266h6cljbr3lr6s3xkrztmlq

Accelerating Cross-correlation Applications via Parallel Computing

M.I. Khalil
2013 International Journal of Image Graphics and Signal Processing  
Moreover, the algorithm portion that is sent to a remote PC, within the LA N, will be redistributed over the available CPU cores on that computer yield ing to maximu m utilizat ion of all available cores  ...  Cross-correlation applicat ions are mostly algorithms rather than data-intensive (that is, they are more CPUbound than I/O-bound).  ...  NET Remoting API The .NET Remot ing API is the equivalent o f the Java Remote Method Invocation (RMI) API.  ... 
doi:10.5815/ijigsp.2013.12.04 fatcat:ibftycuvvrd2jkf6eyv5zwugoy

Opus: A Coordination Language for Multidisciplinary Applications

Barbara Chapman, Matthew Haines, Piyush Mehrotra, Hans Zima, John Van Rosendale
1997 Scientific Programming  
SDAs can be internally data parallel, providing support for the integration of data and task parallelism as well as nested task parallelism.  ...  and heterogeneous in nature, and thus do not fit well into the data parallel paradigm.  ...  Thus, the Opus runtime system must have a mechanism for redistributing data at method invocation time.  ... 
doi:10.1155/1997/632908 fatcat:f2ki3jso7fax5gvvjgwkuacwzy

Design and Implementation of an Environment for Component-Based Parallel Programming [chapter]

Francisco Heron de Carvalho Junior, Rafael Dueire Lins, Ricardo Cordeiro Corrêa, Gisele Araújo, Chanderlie Freire de Santiago
2007 Lecture Notes in Computer Science  
Motivated by the inadequacy of current parallel programming artifacts, the # component model was proposed to meet the new complexity of high performance computing (HPC).  ...  The class of remote service invocation components is represented by the component type RemoteServiceInvocation[e<:Environment, m<:RSIMode, s<:Service].  ...  Figures 1 and 2 complementarily present a simple parallel program that is used for exemplifying the idea of slicing processes by concerns. Let A and B be n×n matrixes and X and Y be vectors.  ... 
doi:10.1007/978-3-540-71351-7_15 fatcat:rrw6sa25h5gjtmqbkozpcjd4ma

Empirical Performance-Model Driven Data Layout Optimization [chapter]

Qingda Lu, Xiaoyang Gao, Sriram Krishnamoorthy, Gerald Baumgartner, J. Ramanujam, P. Sadayappan
2005 Lecture Notes in Computer Science  
The effectiveness of the approach is demonstrated through experimental measurements on some representative computations from quantum chemistry.  ...  The performance model with empirically determined cost components is used to perform data layout optimization in the context of the Tensor Contraction Engine, a compiler for a high-level domainspecific  ...  We also plan to evaluate our approach with other parallel matrix multiplication algorithms.  ... 
doi:10.1007/11532378_7 fatcat:rsnop6hjx5fltjrj2ivbfx2mdi

MADNESS: A Multiresolution, Adaptive Numerical Environment for Scientific Simulation

Robert J. Harrison, Gregory Beylkin, Florian A. Bischoff, Justus A. Calvin, George I. Fann, Jacob Fosso-Tande, Diego Galindo, Jeff R. Hammond, Rebecca Hartman-Baker, Judith C. Hill, Jun Jia, Jakob S. Kottmann (+15 others)
2016 SIAM Journal on Scientific Computing  
Underpinning the numerical capabilities is a powerful petascale parallel programming environment that aims to increase both programmer productivity and code scalability.  ...  numerical environment for scientific simulation) is a high-level software environment for solving integral and differential equations in many dimensions that uses adaptive and fast harmonic analysis methods  ...  The MADNESS runtime spawns a remote method invocation (RMI) thread and a pool of computational threads managed by a task queue; these threads are not directly accessible to the user.  ... 
doi:10.1137/15m1026171 fatcat:sfl7bb7xojg25kkc3kkue4mxge

Formalising Java RMI with explicit code mobility

Alexander Ahern, Nobuko Yoshida
2005 Proceedings of the 20th annual ACM SIGPLAN conference on Object oriented programming systems languages and applications - OOPSLA '05  
Our language captures crucial but often hidden aspects of distributed object-oriented programming, including object serialisation, dynamic class downloading and remote method invocation.  ...  Method invocation Unlike centralised formalisms, DJ describes remote method invocation.  ...  Both remote and local invocations apply this rule. The auxiliary function mbody(m,C, CT) looks up the correct method body in the local class table.  ... 
doi:10.1145/1094811.1094843 dblp:conf/oopsla/AhernY05 fatcat:qfqhxuece5h6zmiilcloqz4jmq

Empirical performance model-driven data layout optimization and library call selection for tensor contraction expressions

Qingda Lu, Xiaoyang Gao, Sriram Krishnamoorthy, Gerald Baumgartner, J. Ramanujam, P. Sadayappan
2012 Journal of Parallel and Distributed Computing  
We observe that the performance of the GEMM routines is significantly influenced by the choice of parameters used in their invocation.  ...  We determine the layouts of the intermediate arrays and the choice of parameters to the GEMM invocations that minimize the overall execution time.  ...  M[ 1 , 1 1], M[2, 1], . . . , M[n, 1], M[1, 2], . . . , M[n, 2], . . . , M[1, m], M[2, m], . . . , M[n, m] The four-dimensional tensor S [a, b, i, j] has the following layout in memory: S [1, 1, 1, 1],  ... 
doi:10.1016/j.jpdc.2011.09.006 fatcat:xxgixibzbzhr3o6sw34sympcpa

Improving communication scheduling for array redistribution

Minyi Guo, Yi Pan
2005 Journal of Parallel and Distributed Computing  
Many scientific applications require array redistribution when the programs run on distributed memory parallel computers.  ...  Thus, the communication idle time is reduced in redistribution routines.  ...  In order to reduce the number of remote accesses, efficient data redistribution is necessary.  ... 
doi:10.1016/j.jpdc.2004.12.001 fatcat:csvulniqdbhqbej6n7tkzzxwoy

Extending Sledgehammer with SMT Solvers

Jasmin Christian Blanchette, Sascha Böhme, Lawrence C. Paulson
2013 Journal of automated reasoning  
We extended Sledgehammer to invoke satisfiability modulo theories (SMT) solvers as well, exploiting its relevance filter and parallel architecture.  ...  The smt Proof Method SMT solvers are available in Isabelle through the smt proof method [15, 18] .  ...  On an N-core machine, Sledgehammer runs at most N provers locally and up to four of the remaining provers remotely.  ... 
doi:10.1007/s10817-013-9278-5 fatcat:nn4vl6f5nrc5pk7opli3l3tmtq

Extending Sledgehammer with SMT Solvers [chapter]

Jasmin Christian Blanchette, Sascha Böhme, Lawrence C. Paulson
2011 Lecture Notes in Computer Science  
We extended Sledgehammer to invoke satisfiability modulo theories (SMT) solvers as well, exploiting its relevance filter and parallel architecture.  ...  The smt Proof Method SMT solvers are available in Isabelle through the smt proof method [15, 18] .  ...  On an N-core machine, Sledgehammer runs at most N provers locally and up to four of the remaining provers remotely.  ... 
doi:10.1007/978-3-642-22438-6_11 fatcat:gv5vgvpccjdarnzc4x7g32dsny

A Survey: Runtime Software Systems for High Performance Computing

2017 Supercomputing Frontiers and Innovations  
These methods they are principally coarse grained and statically scheduled.  ...  The objects communicate through invocation of entry methods. These methods may be called asynchronously and remotely.  ...  Both object instantiation as well as method call is mediated through proxy objects, or lightweight handles representing remote chares.  ... 
doi:10.14529/jsfi170103 fatcat:yqj65kpvhngovcmgrr46vwwr6i

Scilab to Scilab//: The Ouragan project

E Caron, S Chaumette, S Contassot-Vivier, F Desprez, E Fleury, C Gomez, M Goursat, M Quinson, E Jeannot, D Lazure, F Lombard, J.-M Nicod (+7 others)
2001 Parallel Computing  
Mots-clé : SCILAB¢ £ ¢ , bibliothèques parallèles, serveurs de calcul, CORBA, redistribution de données.  ...  In this paper, we present the developments realized in the OURAGAN project around the parallelization of a MATLAB-like tool called SCILAB.  ...  -M. Lepine, F. Pellegrini and N. Viollet.  ... 
doi:10.1016/s0167-8191(01)00101-6 fatcat:lt6yptt3fnherp3sta3qnk54u4

Early-stage performance modeling and its application for integrated embedded control software design

Shige Wang, Kang G. Shin
2004 Software engineering notes  
To demonstrate the usefulness of the proposed method in real world applications, we present the analysis of automotive vehicle-to-vehicle control software as an example.  ...  In this paper, we use a simple heuristic algorithm to find a best-case scenario: if a parallel remote path with the maximum weight (fully parallel) has greater weight than a path of merging the remote  ...  Similarly, a connection < x,y > is said to be asynchronous if the invocations of the action associated with port y do not depend on the output from port x.  ... 
doi:10.1145/974043.974061 fatcat:aspotg3vi5da5h672d5sw4bmf4
« Previous Showing results 1 — 15 out of 989 results