### Citations

56 | Large mesh simplification using processing sequences
- Isenburg, Lindstrom, et al.
- 2003
(Show Context)
Citation Context ...ives many GB data size. There exist special approaches to processing such large meshes, which sizes often exceeds 4GB address space of 32-bit machines. Those approaches may be divided into few groups =-=[10]-=-: • mesh cutting - original mesh is cut into smaller parts, small enough to fit entirely in memory. Then algorithm analyses each part independently. Despite the simplicity of this approach, the initia... |

37 | The Design And Implementation Of The Parallel Out-Of-Core Scalapack
- D’Azevedo, Dongarra
- 2000
(Show Context)
Citation Context ...r as b. As we can see complexity of this algorithm is O(n log n). The IO complexity is O( n b log n k ). By choosing right k we can IO commub nication. 2.2 OOC LU and QR factorization As described in =-=[8]-=- out-of-core extension of ScaLAPACK exists which provides three core factorization routines: LU, QR and Cholesky. They allow factorization of dense systems which are too large to fit in psychical memo... |

27 | A.: Out-of-core compression and decompression of large n-dimensional scalar fields
- IBARRIA, LINDSTROM, et al.
(Show Context)
Citation Context ...proposed in recent years ie. wavelets, Discrete Cosine Transform or Run Length Encoding. Recent research showed that there exists another algorithm, suitable for arbitrary n-dimensional scalar fields =-=[9]-=-. It’s based on the Lorenzo predictor. It estimates the scalar value of a sample on the corner of an n-dimensional cube from the scalar values of the others 2 n − 1 corners. The value of the scalar fi... |

1 |
Challenge Applications. World Wide Web, http://www.mat.univie.ac.at/˜neum/contrib/grand challenges.html
- Grand
(Show Context)
Citation Context ...ored in the memory in the one moment and could be quite big in our opinion. For example there is a class of problems called Grand Challenge Applications which need about 100 gigabytes of data per run.=-=[3]-=- Thus it’s obviously needed to fetch some data into the memory leaving the rest in the memory, process it as much as we can and then take another chunk from mass storage medias ( RAID matrices or simi... |

1 |
library homepage. World Wide Web
- Panda
- 2004
(Show Context)
Citation Context ...s[4]. • Programming libraries which are making programs working more efficiently by giving them good I/O access. It is for example Panda ( daily used in Centre for the Simulation of Advanced Rockets )=-=[5]-=-. These are also lip and KeLPIO ( de5sscribed later ). 2 Problems and solving examples 2.1 Out of core sorting http://www.ds14.agh.edu.pl/∼kubek2k/ooc/ Sorting is a fundamental procedure, useful in ma... |

1 | Grid Generation and Manipulation Tools. www.cerfacs.fr/˜muller/delaundo.html. Title image - Muller |

1 | Collection of quotations - Stepney - 2004 |

1 |
lip library homepage. World Wide Web, http://galaxy.uci.agh.edu.pl/˜kzajac/, 2002. lip documentation
- Zajac
(Show Context)
Citation Context ...ssion method which naturally generate triangles and vertex ordering suitable as the processing sequence. 3 OOC environments 3.1 lip environment One of OOC libraries on which I want to concern is lip. =-=[13]-=- It was made as a utility able to help in solving Irregular 3 simulations and Out of Core problems. It is based on MPI-IO, so it can be ported on any system: from SMP machine to simple workstation. It... |

1 |
On-line OOC tutorial. World Wide Web, http://galaxy.uci.agh.edu.pl/˜kzajac/ooc tut/, 2004. Part of lip documentation
- Zajac
(Show Context)
Citation Context ...solving 2 technologically-achievable :P 2s1.2 Variation of abstract models http://www.ds14.agh.edu.pl/∼kubek2k/ooc/ There are three most common kinds of OOC’s solve models used in parallel computing: =-=[14]-=- • General placement model – when the memory is available for all the processes and is stored in one BIG virtual file – figure 1 Proccess 1 One file for all processess Process 2 Fig. 1: General placem... |