Data-Parallel Implementations of Dense Simplex Methods on the Connection Machine CM-2

We study numerically the growth of a crack in an elastic medium under the influence of a travelling shockwave. We describe the implementation of a fast algorithm which is perfectly suited for a data parallel computer. Using large scale simulations on the Connection Machine we generate cracks with more than 10000 sites on a 1024 × 1024 lattice. We show that the resulting patterns are fractal with a fractal dimension that depends on the chosen breaking criterion and varies between 1. and 2.

Download Full-text

MASSIVELY DATA-PARALLEL MOLECULAR DYNAMICS

International Journal of Modern Physics C ◽

10.1142/s0129183192000440 ◽

1992 ◽

Vol 03 (04) ◽

pp. 709-731

Author(s):

ERNESTO BONOMI ◽

MARCO TOMASSINI

Keyword(s):

Molecular Dynamics ◽

Heat Bath ◽

Hard Core ◽

Parallel Computers ◽

Particle Systems ◽

Data Parallel ◽

Connection Machine ◽

Large N ◽

Parallel Molecular Dynamics ◽

Dynamics Simulations

In light of present day data-parallel computers, an appraisal of molecular dynamics simulations of large N-particle systems, isolated or in contact with a heat-bath, is given. Special attention is focused 011 the Connection Machine CM-2. Particularly the cases of long-range potentials and impulsive hard-core interactions are discussed in detail. Data-parallel strategies including data distribution, communications and computation are presented and compared with well-known sequential approaches. The conclusion offered is that the methods described here are easy to design and offer the possibility of reasonably fast implementations for the reliable simulation of macroscopic samples of matter.

Download Full-text

STEALTH TERRAIN NAVIGATION FOR MULTI-VEHICLE PATH PLANNING

International Journal of Pattern Recognition and Artificial Intelligence ◽

10.1142/s0218001492000242 ◽

1992 ◽

Vol 06 (02n03) ◽

pp. 395-415 ◽

Cited By ~ 1

Author(s):

Y. ANSEL TENG ◽

DANIEL DEMENTHON ◽

LARRY S. DAVIS

Keyword(s):

Path Planning ◽

Communication Complexity ◽

Parallel Machines ◽

Solution Space ◽

Planning System ◽

Data Parallel ◽

Temporal Sampling ◽

Connection Machine ◽

Vehicle Path ◽

Using Data

In this paper, we propose a method for solving visibility-based terrain path planning problems for groups of vehicles using data parallel machines. The discussion focuses on path planning for two groups of vehicles so that they move in a bounding overwatch manner. Furthermore, the planned paths for the vehicles themselves are subject to intervisibility constraints, configuration constraints, and different terrain traversabilities due to variations in terrain type and slope. A spatial-temporal sampling approach is adopted to discretize the solution space and facilitate fast computation on a data parallel machine. One of the key computations in the planning is the region-to-region visibility analysis, which is computationally expensive but essential to the choice of subgoals to carry out reconnaissance activities. A parallel algorithm for this analysis is developed. By reducing the communication complexity, our algorithm achieves much faster running time than traditional methods. The algorithms are implemented on a Connection Machine CM-2, and the experimental results show that the planning system effectively generates good paths.

Download Full-text

DYNAMIC LOAD-BALANCING STRATEGIES FOR DATA PARALLEL IMPLEMENTATIONS OF REACTION-EVOLUTION-MIGRATION SYSTEMS

International Journal of Modern Physics C ◽

10.1142/s0129183193000124 ◽

1993 ◽

Vol 04 (01) ◽

pp. 107-119 ◽

Cited By ~ 1

Author(s):

MARK SMITH

Keyword(s):

Load Balancing ◽

Dynamic Load ◽

Poor Performance ◽

Scientific Field ◽

Stochastic Simulations ◽

Dynamic Load Balancing ◽

Data Parallel ◽

Parallel Implementations ◽

Migration Systems

We introduce Reaction-Evolution-Migration Systems and explain their importance in the scientific field. Details are given of data parallel implementations of such systems, and how a naïve attempt can give very poor performance for stochastic simulations. Thus details are given of a number of dynamic load-balancing strategies to counter this problem, along with comparisons of the performance and ease of implementation of these strategies.

Download Full-text

POMPC: A C LANGUAGE FOR DATA PARALLELISM

International Journal of Modern Physics C ◽

10.1142/s0129183193000094 ◽

1993 ◽

Vol 04 (01) ◽

pp. 85-96

Author(s):

NICOLAS PARIS

Keyword(s):

Parallel Computers ◽

Massively Parallel ◽

C Language ◽

Parallel Language ◽

Data Parallel ◽

Massively Parallel Computers ◽

Connection Machine ◽

Compilation Process ◽

Points Of View ◽

Definition Of

POMPC is a parallel language dedicated to the programming of Massively Parallel Computers according to a synchronous Data Parallel model. It is an extension of the ANSI C language and follows its philosophy. Parallelism is explicitly handled by the definition of collections of parallel variables and the definition of communication primitives. A methodology is presented in order to easily port the language on different target architectures. Virtualization is introduced to handle simultaneously several collections of different sizes and shapes. Virtualization management is a key point of the compilation process. Programmer, architecture, compilation and system points of view lead to a special implementation of the virtualization mixing physical and virtual parallel objects. The implementation of the virtualization is adapted for the development of communication libraries and also adapted to enlarge the asynchronous sections of code for SPMD architecture. The portability of the POMPC language is validated by several implementations for mono/multi-process simulation on UNIX machines, for the Connection Machine CM-2, for the MasPar MP-1 and a compiler is in preparation for the iPSC-860.

Download Full-text