## A.: An object-oriented platform for distributed high-performance Symbolic Computation (1999)

Venue: | Mathematics and Computers in Simulation 49 |

Citations: | 17 - 11 self |

### BibTeX

@INPROCEEDINGS{Blochinger99a.:an,

author = {Wolfgang Blochinger and Wolfgang Küchlin and Christoph Ludwig and Andreas Weber},

title = {A.: An object-oriented platform for distributed high-performance Symbolic Computation},

booktitle = {Mathematics and Computers in Simulation 49},

year = {1999},

pages = {161--178}

}

### OpenURL

### Abstract

We describe the Distributed Object-Oriented Threads System (DOTS), a programming environment designed to support object-oriented fork/join parallel programming in a heterogeneous distributed environment. A mixed network of Windows NT PC’s and UNIX workstations is transformed by DOTS into a homogeneous pool of anonymous compute servers forming together a multicomputer. DOTS is a complete redesign of the Distributed Threads System (DTS) using the object-oriented paradigm both in its internal implementation and in the programming paradigm it supports. It has been used for the parallelization of applications in the field of computer algebra and in the field of computer graphics. We also give a brief account of applications in the domain of symbolic computation that were developed using DTS. Key words: distributed threads system, heterogeneous networks, Windows NT

### Citations

913 |
A Course in Computational Algebraic Number Theory
- Cohen
- 1993
(Show Context)
Citation Context ...been developed, many of them during the last twenty years. The Elliptic Curve Method. H. W. Lenstra’s Elliptic Curve Method (ECM) [20], one of these algorithms, advances on the idea of the p−1-method =-=[24,7]-=-, but it replaces the residues modN by points on elliptic curves modN. The points on an elliptic curve with coordinates over a field of characteristic p �∈ {2, 3} form an Abelian group. However, the g... |

250 |
Algorithms for Computer Algebra
- Geddes, Czapor, et al.
- 1992
(Show Context)
Citation Context ... ,xr−1 are such that R(x1,... ,xr−1) is zero. Resultants have important applications in Computer Algebra, such as performing exact arithmetic on algebraic numbers [22] or solving systems of equations =-=[13]-=-. The modular method [8] is a divide-and-conquer scheme which recursively maps the multivariate resultant computation to multiple resultant computations of homomorphic images, and, using the Chinese R... |

233 |
Factoring Integers with Elliptic Curves
- Lenstra
- 1987
(Show Context)
Citation Context ...tography. Therefore a lot of algorithms approaching this problem have been developed, many of them during the last twenty years. The Elliptic Curve Method. H. W. Lenstra’s Elliptic Curve Method (ECM) =-=[20]-=-, one of these algorithms, advances on the idea of the p−1-method [24,7], but it replaces the residues modN by points on elliptic curves modN. The points on an elliptic curve with coordinates over a f... |

207 | The Nexus approach to integrating multithreading and communication
- Foster, Kesselman, et al.
- 1996
(Show Context)
Citation Context ...System Hardware Network Fig. 11. The Distributed PaReDuX system [6]. PaReDuX system, cf. [6]. The architecture of the Distributed PaReDuX system is sketched in Fig. 11. 6 Related Work 6.1 Nexus Nexus =-=[10,11]-=- provides an integration of multi-threading and communication by separating the specification of the communication’s destination and the specification of the thread of control that should respond to t... |

186 |
Speeding the Pollard and elliptic curve methods of factorization
- Montgomery
- 1987
(Show Context)
Citation Context ...been developed, many of them during the last twenty years. The Elliptic Curve Method. H. W. Lenstra’s Elliptic Curve Method (ECM) [20], one of these algorithms, advances on the idea of the p−1-method =-=[24,7]-=-, but it replaces the residues modN by points on elliptic curves modN. The points on an elliptic curve with coordinates over a field of characteristic p �∈ {2, 3} form an Abelian group. However, the g... |

108 | The ADAPTIVE Communication Environment: An object-oriented network programming toolkit for developing communication software,” Concurrency
- Schmidt
- 1993
(Show Context)
Citation Context ...rk has the advantage that the argument object is frozen at the time of the fork. 3.4 Implementation of DOTS The implementation of DOTS is based on the Adaptive Communication Environment (ACE) toolkit =-=[28]-=-. ACE is a system-independent, object-oriented platform for developing communication software. There are implementations 9sDOTS ACE OS ACE SOCK SAP Log Msg UNIX Log Msg NT ACE FIFO ACE SPIPE ACE SAP S... |

80 |
The GNU multiple precision arithmetic library. http://www.swox.com/gmp
- Granlund
(Show Context)
Citation Context ...e. Therefore the distributed version of the ECM had to be re-implemented, but the code follows LiDIA closely. The Implementation. Since there were already positive experiences with the GNU MP library =-=[15]-=- in multi-threaded programs—one only has to apply a minor patch—the new implementation relies on this library for the involved 11sarithmetic. The element methods of the residue classes essentially exp... |

64 | Adaptive and reliable parallel computing on networks of workstations
- BLUMOFE, L
- 1997
(Show Context)
Citation Context ...lel programming language compilers, whereas DOTS is intended to be used directly by the application programmer. 19s6.2 Cilk-NOW Cilk is a parallel multi-threaded extension of the C language. Cilk-NOW =-=[1]-=- provides a runtime-system for a functional subset of Cilk that enables the user to run Cilk programs on networks of UNIX workstations. Cilk-NOW supports multi-threading in a continuation passing styl... |

52 | ObjectOriented Components for High-speed Network Programming
- Schmidt, Harrison, et al.
- 1995
(Show Context)
Citation Context ...improves the type-safety of the implementation. This can prevent the occurrence of run time errors and therefore improves the correctness and software quality. • Performance measurements presented in =-=[29]-=- show that the additional abstraction layers of ACE do not cause any significant performance loss compared with using operating system calls directly. 4 Applications and Performance Measurements This ... |

51 |
The PVM system: Supercomputing level concurrent computations on a heterogeneous network of workstations
- Geist, Sunderam
- 1991
(Show Context)
Citation Context ...the domain of computer graphics— this issue will be one focus in our future research activities. 5 DTS The Distributed Threads System DTS [2] is the ancestor of DOTS. It was implemented on top of PVM =-=[14]-=- and is well suited for the parallelization of C based applications. Parameter marshalling in DTS is achieved via user supplied copying functions, while in DOTS object serialization is used. In DTS, s... |

44 | The Nexus Task-parallel Runtime System
- Foster, Kesselamn, et al.
- 1994
(Show Context)
Citation Context ...System Hardware Network Fig. 11. The Distributed PaReDuX system [6]. PaReDuX system, cf. [6]. The architecture of the Distributed PaReDuX system is sketched in Fig. 11. 6 Related Work 6.1 Nexus Nexus =-=[10,11]-=- provides an integration of multi-threading and communication by separating the specification of the communication’s destination and the specification of the thread of control that should respond to t... |

37 |
The calculation of multivariate polynomial resultants
- COLLINS
- 1971
(Show Context)
Citation Context ...,... ,xr−1) is zero. Resultants have important applications in Computer Algebra, such as performing exact arithmetic on algebraic numbers [22] or solving systems of equations [13]. The modular method =-=[8]-=- is a divide-and-conquer scheme which recursively maps the multivariate resultant computation to multiple resultant computations of homomorphic images, and, using the Chinese Remainder Algorithm, lift... |

28 |
Computing in algebraic extensions
- Loos
- 1983
(Show Context)
Citation Context .... (and hence common roots), if x1,... ,xr−1 are such that R(x1,... ,xr−1) is zero. Resultants have important applications in Computer Algebra, such as performing exact arithmetic on algebraic numbers =-=[22]-=- or solving systems of equations [13]. The modular method [8] is a divide-and-conquer scheme which recursively maps the multivariate resultant computation to multiple resultant computations of homomor... |

20 |
Millipede: Easy Parallel Programming in Available Distributed Environments, Software: Practice and Experience 27(8), 929–965
- FRIEDMAN, GOLDIN, et al.
- 1997
(Show Context)
Citation Context ...ments a work-steeling scheduling technique. In DOTS the Global Manager can be configured with application dependent schedulers by subclassing the system’s scheduler interface. 6.3 Millipede Millipede =-=[12]-=- is a system that also focuses on providing a parallelization environment on widely available hardware platforms. The essential goal of Millipede is to provide a distributed shared memory (DSM) enviro... |

20 |
PARSAC-2: A parallel SAC-2 based on threads
- Küchlin
- 1991
(Show Context)
Citation Context ... DOTS has adopted many of the basic concepts introduced by its predecessor system DTS [2]. The primary motivation for the development of DTS was to extend the target platforms of the PARSAC-2 library =-=[17,18]-=- from shared memory multiprocessors to a network of shared memory multiprocessors. Besides being suitable for the realization of this initial goal, it turned out that the distributed fork/join paradig... |

16 | Distributed symbolic computation with DTS
- Bubeck, Hiller, et al.
- 1995
(Show Context)
Citation Context ...ct serialization of the parameter and result objects that are involved in the distributed part of the computation. DOTS has adopted many of the basic concepts introduced by its predecessor system DTS =-=[2]-=-. The primary motivation for the development of DTS was to extend the target platforms of the PARSAC-2 library [17,18] from shared memory multiprocessors to a network of shared memory multiprocessors.... |

13 |
A C++ library for computational number theory. http://www.informatik.tu-darmstadt.de/TI/LiDIA
- LIDIA
(Show Context)
Citation Context ...f additions of points (on several curves) one is willing to perform, one can control the probability that the algorithm finds a nontrivial divisor of N. LiDIA. The computational algebra library LiDIA =-=[21]-=- contains a sequential implementation of the ECM that uses the suggestions in [24]. This algorithm performs the group law on a number of elliptic curves, and the calculation on different curves are co... |

10 |
A master-slave approach to parallel term rewriting on a hierarchical multiprocessor”, Lecture
- Bündgen, Göbel, et al.
(Show Context)
Citation Context ... of projects whose aim was to speed up symbolic computations in both Computer Algebra and Computational Logic by parallelization on a network of (possibly parallel) workstations—e. g. theorem provers =-=[4,6]-=-, quantifier elimination packages [9], complex root finders [27], or symbolic solvers [2,25]. 5.1 Parallel Resultant Computation in PARSAC-2 Given two r-variate polynomials P, Q ∈ Z[x1,... ,xr], we ar... |

10 |
The S-Threads Environment for Parallel Symbolic Computation
- Küchlin
(Show Context)
Citation Context ... marshalling in DTS is achieved via user supplied copying functions, while in DOTS object serialization is used. In DTS, special attention was directed to transparently integrate the S-threads system =-=[16]-=- for shared-memory parallel symbolic computation into a distributed environment. DTS has been used successfully for a variety of projects whose aim was to speed up symbolic computations in both Comput... |

10 |
S/390 Cluster Technology: Parallel Sysplex
- Nick
- 1997
(Show Context)
Citation Context ...ster of Windows NT workstations. 7 Future Work One of our future goals is to support a wider range of platforms. We are planning to provide a version of DOTS for the IBM S/390 Parallel Sysplex system =-=[26]-=-, a clustered multiprocessor architecture. DOTS will then be available on a wide range of system platforms including low cost Windows NT PC networks and high end mainframe clusters, providing one sing... |

7 | PARSAC-2: Parallel computer algebra on the desk-top
- Kuchlin
- 1995
(Show Context)
Citation Context ... DOTS has adopted many of the basic concepts introduced by its predecessor system DTS [2]. The primary motivation for the development of DTS was to extend the target platforms of the PARSAC-2 library =-=[17,18]-=- from shared memory multiprocessors to a network of shared memory multiprocessors. Besides being suitable for the realization of this initial goal, it turned out that the distributed fork/join paradig... |

6 |
Parallel ReDuX → PaReDuX
- Bündgen, Göbel, et al.
- 1995
(Show Context)
Citation Context ...ased on an unfailing completion procedure, on a network of multiprocessor SPARC-stations. The functions, which are distributed over the network via DTS, are implemented in the PaReDuX system. PaReDuX =-=[5,3]-=- is a multi-threaded parallel equational theorem prover based on the S-threads package. It provides several so called strategy compliant completion algorithms. These work on each multi-processor node.... |

6 |
Parallel direct volume rendering on PC networks
- MEISSNER, HÜTTNER, et al.
- 1998
(Show Context)
Citation Context ... power, the fast rendering of 3D volumetric data becomes more and more crucial for many applications in the field of medical science and in other scientific domains. As is described in full detail in =-=[23]-=-, DOTS was successfully used to considerably speed up a volume rendering application on inexpensive and widely used Windows NT PCs. The pool consisted of up to five Intel PentiumPro and up to four Int... |

5 | Parallel Term Rewriting with PaReDuX
- Bündgen, Göbel, et al.
- 1998
(Show Context)
Citation Context ... of projects whose aim was to speed up symbolic computations in both Computer Algebra and Computational Logic by parallelization on a network of (possibly parallel) workstations—e. g. theorem provers =-=[4,6]-=-, quantifier elimination packages [9], complex root finders [27], or symbolic solvers [2,25]. 5.1 Parallel Resultant Computation in PARSAC-2 Given two r-variate polynomials P, Q ∈ Z[x1,... ,xr], we ar... |

4 |
Strategy compliant multi-threaded term completion
- Bündgen, Göbel, et al.
- 1996
(Show Context)
Citation Context ...ased on an unfailing completion procedure, on a network of multiprocessor SPARC-stations. The functions, which are distributed over the network via DTS, are implemented in the PaReDuX system. PaReDuX =-=[5,3]-=- is a multi-threaded parallel equational theorem prover based on the S-threads package. It provides several so called strategy compliant completion algorithms. These work on each multi-processor node.... |

3 | Approaches to parallel quantifier elimination
- Dolzmann, Gloor, et al.
- 1998
(Show Context)
Citation Context ...symbolic computations in both Computer Algebra and Computational Logic by parallelization on a network of (possibly parallel) workstations—e. g. theorem provers [4,6], quantifier elimination packages =-=[9]-=-, complex root finders [27], or symbolic solvers [2,25]. 5.1 Parallel Resultant Computation in PARSAC-2 Given two r-variate polynomials P, Q ∈ Z[x1,... ,xr], we are interested in their common roots. T... |

2 | A parallel complex zero finder
- SCHAEFER, BUBECK
- 1995
(Show Context)
Citation Context ...oth Computer Algebra and Computational Logic by parallelization on a network of (possibly parallel) workstations—e. g. theorem provers [4,6], quantifier elimination packages [9], complex root finders =-=[27]-=-, or symbolic solvers [2,25]. 5.1 Parallel Resultant Computation in PARSAC-2 Given two r-variate polynomials P, Q ∈ Z[x1,... ,xr], we are interested in their common roots. The resultant R(P, Q) is a p... |

1 |
Verteiltes Berechnen von Gröbner Basen mit S-Threads und DTS. Diplomarbeit, Universität Tübingen, Fakultät für Informatik
- Nagel
- 1997
(Show Context)
Citation Context ...omputational Logic by parallelization on a network of (possibly parallel) workstations—e. g. theorem provers [4,6], quantifier elimination packages [9], complex root finders [27], or symbolic solvers =-=[2,25]-=-. 5.1 Parallel Resultant Computation in PARSAC-2 Given two r-variate polynomials P, Q ∈ Z[x1,... ,xr], we are interested in their common roots. The resultant R(P, Q) is a polynomial in r − 1variables ... |