Segmentation and other image processing operations rely on convolution calculations with heavy computational and memory access demands. This paper presents an analysis of a texture segmentation application containing a 96x96 convolution. Sequential execution required several hours on single processor systems with over 99% of the time spent performing the large convolution. 70% to 75% of executi...

By viewing di erent parallel programming paradigms as essential heterogeneous approaches in mapping \real-world" problems to parallel systems, we discuss methodologies in integrating multiple programming models on a Connection Machine CM5. In a data ow based integration model built in a visualization software AVS, we demonstrate a simple, e ective and modular way to couple sequential, data-para...

Most existing inductive learning systems form concept descriptions in propositional languages from vectors of basic features. However, many concepts are characterized by the relationships of individual examples to general domain knowledge. We describe a system that constructs relational terms e~ciently to augment the description language of standard inductive systems. In our approach, examples ...

We consider the parallelization of two standard 2D reconstruction algorithms, ®ltered backprojection and direct Fourier reconstruction, using the data-parallel programming style. The algorithms are implemented on a Connection Machine CM-5 with 16 processors and a peak performance of 2 G ̄op/s. Ó 1998 Elsevier Science B.V. All rights reserved.

We give two parallel algorithms for sequence comparison on the Connection Machine 2 (CM2). The specific comparison measure we compute is the erfif distance: given a finite alphabet £ and two input sequences X £ £ + and Y £ E + the edit distance Y ) is the minimum cost of transforming X into Y via a series of weighted insertions, deletions, and substitutions of characters. The edit distance comp...

By viewing diierent parallel programming paradigms as essential heterogeneous approaches in mapping \real-world" problems to parallel systems, we discuss method-ologies in integrating multiple programming models on a Connection Machine CM5. In a dataaow based integration model built in a visualization software AVS, we demonstrate a simple, eeective and modular way to couple sequential, data-par...

A fast algorithm for the fault simulation using data level parallelism is implemented on the Connection Machine. The algorithm employed is the PPSFP (Paralle l Pattern Single Fault Propagation) type. The input of the algorithm is the PI (Primary Input) test patterns obtained from a cellular automata pseudo-random pattern generator. The output of this system is a fault simulation table that desc...

An efficient unstructured 3D Euler solver is parallelized on a Thinking Machine Corporation Connection Machine 5, distributed memory computer with vectorizing capability. In this paper, the SIMD strategy is employed through the use of the CM Fortran language and the CMSSL scientific library. The performance of the CMSSL mesh partitioner is evaluated and the overall efficiency of the parallel fl...