Filters








11,958 Hits in 3.1 sec

A semantic framework to address data locality in data parallel languages

Eric Violard
2004 Parallel Computing  
Consequently these two languages define their own balance between compiler and programmer investments in order to reach program efficiency.  ...  This theory deals with data locality which is a main issue in parallel programming. In this article, we regard this theory and its model as a minimum semantic domain for data parallel languages.  ...  variables'' and ''data locality'', while the compiler can be left in charge of the distribution onto physical processors, communications management, etc.  ... 
doi:10.1016/s0167-8191(03)00089-9 fatcat:etx2jb262zardj3uv3x327e3oa

Task-Oriented Vision [chapter]

Katsushi Ikeuchi, Martial Hebert
1996 Exploratory Vision  
Based on this discussion, we will develop the task-oriented vision paradigm, and examine the difference between the task-oriented vision paradigm and the traditional Marr's paradigm.  ...  We will also explore the research issues necessary for completing the taskoriented vision paradigm.  ...  The cenual issue in the research is how to build the compiler, which automatically converts a geometric and sensor model into a localization program.  ... 
doi:10.1007/978-1-4612-3984-0_11 fatcat:oqdk6da7hfb7xaqjlmug3dzugi

Evaluation of PGAS Communication Paradigms with Geometric Multigrid

Hongzhang Shan, Amir Kamil, Samuel Williams, Yili Zheng, Katherine Yelick
2014 Proceedings of the 8th International Conference on Partitioned Global Address Space Programming Models - PGAS '14  
In this paper, we evaluate three different one-sided communication paradigms in the context of geometric multigrid using the miniGMG benchmark.  ...  Partitioned Global Address Space (PGAS) languages and one-sided communication enable application developers to select the communication paradigm that balances the performance needs of applications with  ...  However, the fine-grained communication paradigm enables faster application development with less code, and performance can always be improved through incremental optimizations.  ... 
doi:10.1145/2676870.2676874 dblp:conf/pgas/ShanKWZY14 fatcat:wmszvqh4ojcsnaqvywcu6w5dlm

Data Parallel Iterators for Hierarchical Grid and Tree Algorithms [chapter]

Gerhard Zumbusch
2006 Lecture Notes in Computer Science  
For a full compile time dependence analysis and a tight distributed memory parallelization, some additional application knowledge about alignment of arrays or indirect data access can be put into the application's  ...  The "owner computes" paradigm guarantees local memory store operations, such that non-local load operations are the main source of inter process communication.  ...  Hence, the application programmer provides the geometric interpretation implicitly for parallelization through the specification of a block distribution.  ... 
doi:10.1007/11823285_65 fatcat:xlbqhge2crct3g7iujnmb4wmuu

Scalable Large Scale Process Modeling and Simulations in Liquid Composite Molding [chapter]

Ram Mohan, Dale Shires, Andrew Mark
2001 Lecture Notes in Computer Science  
These approaches continue to mature through more robust compilers and an enhanced standard.  ...  In the MPI based paradigms, parallelism is obtained through a single-program, multiple-data (SPMD) programming model.  ... 
doi:10.1007/3-540-45545-0_131 fatcat:2ywmjsleefhmxfl7tplkwh6njq

Processor Scheduling On Parallel Computers

Mohammad S. Laghari, Gulzar A. Khuwaja
2012 Zenodo  
Many problems in computer vision and image processing present potential for parallel implementations through one of the three major paradigms of geometric parallelism, algorithmic parallelism and processor  ...  This paper considers the application of parallel or multi-computers to a class of problems exhibiting spatial data characteristic of the geometric paradigm.  ...  Fig. 1 (a) SISD computer Fig. 2 2 Example algorithmic parallelism -a language compiler Fig. 3 3 Fig. 3 Example geometric structure Fig. 5 . 5 Fig. 5 5 Computer network for the topology Fig. 6  ... 
doi:10.5281/zenodo.1085079 fatcat:5qtqt32fevhpppglshyv5mosnu

From swarm robotics to smart materials

Nikolaus Correll, Roderich Groß
2010 Neural computing & applications (Print)  
, and local communication.  ...  Proto code is then compiled into op-codes for the Proto Virtual Machine, which needs to provide abstractions for sensing, actuation, estimation of the geometric relations between neighboring swarm members  ...  , and local communication.  ... 
doi:10.1007/s00521-010-0440-2 fatcat:36ndlvdltfeetjyqmevrburhvu

Parallelization of a dynamic unstructured algorithm using three leading programming paradigms

L. Oliker, R. Biswas
2000 IEEE Transactions on Parallel and Distributed Systems  
This paper presents the parallelization of a dynamic unstructured mesh adaptation algorithm using three popular programming paradigms on three leading supercomputers.  ...  The MTA code was compiled on the MTX 0.5.1 operating system with Tera C version 0.5. The number of streams is easily changed through a compiler directive.  ...  Dynamic distribution of loop iterations among the processors is easily achieved through pragma compiler directives.  ... 
doi:10.1109/71.879776 fatcat:6gjamrhcmrb7biabyr2yguzwrq

Classical computational geometry in GeomNet

Gill Barequet, Stina S. Bridgeman, Christian A. Duncan, Michael T. Goodrich, Roberto Tamassia
1997 Proceedings of the thirteenth annual symposium on Computational geometry - SCG '97  
In this paper we present GeomNet, a system for performing dktributed geometric computing over the Internet.  ...  We also provide seved examples of actual geometric algorithms that our system already supports. Application domains for GeomNet include collaborative research and dist ante education.  ...  standard Java computing paradigm.  ... 
doi:10.1145/262839.263035 dblp:conf/compgeom/BarequetBDGT97 fatcat:ax6ffc37ojeilnmsfqgehmqhya

Performance Analysis of a Hybrid Overset Multi-block Application on Multiple Architectures [chapter]

M. Jahed Djomehri, Rupak Biswas
2003 Lecture Notes in Computer Science  
The application is implemented using a hybrid MPI+OpenMP programming paradigm that exploits both coarse and fine-grain parallelism; the former via MPI message passing and the latter via OpenMP directives  ...  The overset grid application is popular within the aerodynamics community due to its ability to handle complex designs with multiple geometric components, whereby individual body-fitting grids are easily  ...  Since the SXG vector unit is significantly more powerful than the scalar processor. it is critical to achieve high vector operation ratios, either via compiler discovery or explicitly through code (re-  ... 
doi:10.1007/978-3-540-24596-4_41 fatcat:wbl3h6vdgfcnfbkzua7gpyjdz4

Problem-solving Environments For Computational Science

E. Houstis, E. Gallopoulos, R. Bramley, J. Rice
1997 IEEE Computational Science & Engineering  
geometric and physical interfaces.  ...  System software In order for the NC paradigm to be realized, we in the PSE community must standardize both low-level and high-level library interfaces.  ... 
doi:10.1109/mcse.1997.615427 fatcat:d3k65i4rnvfmxhfpdiw56pgema

Parallelization of a dynamic unstructured application using three leading paradigms

Leonid Oliker, Rupak Biswas
1999 Proceedings of the 1999 ACM/IEEE conference on Supercomputing (CDROM) - Supercomputing '99  
This paper presents the parallelization of a dynamic unstructured mesh adaptation algorithm using three popular programming paradigms on three leading supercomputers.  ...  Table 5 shows the performance on the 255 MHz Tera MTA, through five refinement levels. The number of streams is easily changed through a compiler directive.  ...  Dynamic distribution of loop iterations among the processors is easily achieved through pragma compiler directives.  ... 
doi:10.1145/331532.331571 dblp:conf/sc/OlikerB99 fatcat:3fbdskwh3bb3dnpwm73zag4xlm

High performance Fortran for highly irregular problems

Y. Charlie Hu, S. Lennart Johnsson, Shang-Hua Teng
1997 SIGPLAN notices  
Our data parallel N -body code provides a much needed "benchmark" code for evaluating and improving HPF compilers.  ...  for linearizing irregular data structures (2) a data parallel implementation (in HPF) of graph partitioning algorithms applied to the linearized data structure, (3) techniques for expressing irregular communication  ...  Acknowledgment We would like to thank Doug Miles and Paul Kinney for their help with performance tuning of the adaptive code using the pghpf compiler, and Michael Warren for providing the Cold Dark matter  ... 
doi:10.1145/263767.263769 fatcat:dudca3z2qbggnmo4ousx2rog3u

High performance Fortran for highly irregular problems

Y. Charlie Hu, S. Lennart Johnsson, Shang-Hua Teng
1997 Proceedings of the sixth ACM SIGPLAN symposium on Principles and practice of parallel programming - PPOPP '97  
Our data parallel N -body code provides a much needed "benchmark" code for evaluating and improving HPF compilers.  ...  for linearizing irregular data structures (2) a data parallel implementation (in HPF) of graph partitioning algorithms applied to the linearized data structure, (3) techniques for expressing irregular communication  ...  Acknowledgment We would like to thank Doug Miles and Paul Kinney for their help with performance tuning of the adaptive code using the pghpf compiler, and Michael Warren for providing the Cold Dark matter  ... 
doi:10.1145/263764.263769 dblp:conf/ppopp/HuJT97 fatcat:3sevabxntjhmnoixjgmg6oloni

An agent framework for agent societies

Kyle Usbeck, Jacob Beal
2011 Proceedings of the compilation of the co-located workshops on DSM'11, TMC'11, AGERE!'11, AOOPES'11, NEAT'11, & VMIL'11 - SPLASH '11 Workshops  
The spatial computing language Proto, however, compiles descriptions of global behavior into local behaviors that interact to produce the specified emergent behavior.  ...  (Photo credit: (b) James McLurkin and Swaine Photography, (c) Nikolaus Correll, (d) Hayes Raffle and Amanda Parkes) A manifold is a geometric space that locally resembles Euclidean space, but globally  ...  Thus, the messages between Proto software agents are passed through shared memory, often made available by the hardware agent's implied communication.  ... 
doi:10.1145/2095050.2095082 dblp:conf/oopsla/UsbeckB11 fatcat:ec6tssggxvgjfc7unxtjzchej4
« Previous Showing results 1 — 15 out of 11,958 results