## The Multicomputer Toolbox: Scalable Parallel Libraries for Large-Scale Concurrent Applications (1994)

Citations: | 19 - 11 self |

### BibTeX

@TECHREPORT{Skjellum94themulticomputer,

author = {Anthony Skjellum and Chuck Baldwin},

title = {The Multicomputer Toolbox: Scalable Parallel Libraries for Large-Scale Concurrent Applications},

institution = {},

year = {1994}

}

### Years of Citing Articles

### OpenURL

### Abstract

In this paper, we consider what is required to develop parallel algorithms for engineering applications on message-passing concurrent computers (multicomputers). At Caltech, the first author studied the concurrent dynamic simulation of distillation column networks [19, 21, 20, 14]. This research was accomplished with attention to portability, high performance and reusability of the underlying algorithms. Emerging from this work are several key results: first, a methodology for explicit parallelization of algorithms and for the evaluation of parallel algorithms in the distributed-memory context; second, a set of portable, reusable numerical algorithms constituting a "Multicomputer Toolbox," suitable for use on both existing and future medium-grain concurrent computers; third, a working prototype simulation system, Cdyn, for distillation problems, that can be enhanced (with additional work) to address more complex flowsheeting problems in chemical engineering; fourth, ideas for how to a...

### Citations

166 |
Performance Computer Architecture
- Stone
(Show Context)
Citation Context ...tion on two r-vectors of real numbers), the recursive doubling procedure produces the combined result in each of the R processes in dlog 2 Re steps. This procedure is symbolized in Figure 9. See also =-=[22]-=-. Each step involves the cost of locally combining two instances of O to yield O 0 , the cost of transmitting the object O 0 via F , and the cost of receiving another object O 00 . It is crucial that ... |

55 |
Zipcode: A portable multicomputer communication library atop the reactive kemal
- Skjellum, Leung
- 1990
(Show Context)
Citation Context ...l work) to address more complex flowsheeting problems in chemical engineering; fourth, ideas for how to achieve higher performance with Cdyn, using iterative methods for the underlying linear algebra =-=[16, 15, 17, 14]-=-. Of these, the chief emphasis in the present paper is the reusable collection of parallel libraries comprising the Toolbox. Concurrent algorithms for the solution of dense and sparse linear systems, ... |

15 |
de Velde. Data Redistribution and Concurrency
- van
- 1990
(Show Context)
Citation Context ...lization of x to zero at the outset allows for the negligible "bookkeeping" inherent in this conversion procedure. 5.4 Inner Products First we define two conventional inner products per, for=-= example, [23, 24]-=-. Then we consider a new kernel, the skew inner product, and a communication-free kernel, skew outer product. Definition 5.5 (Inner Product) The simplest inner product x T y of two concurrent vectorss... |

13 |
Concurrent Dynamic Simulation: Multicomputer Algorithms Research Applied to Ordinary Di erential-Algebraic
- Skjellum
- 1990
(Show Context)
Citation Context ...lgorithms for engineering applications on message-passing concurrent computers (multicomputers). At Caltech, the first author studied the concurrent dynamic simulation of distillation column networks =-=[19, 21, 20, 14]-=-. This research was accomplished with attention to portability, high performance and reusability of the underlying algorithms. Emerging from this work are several key results: first, a methodology for... |

11 |
Zipcode: A Portable Communication Layer for High Performance Multicomputing
- Skjellum, Smith, et al.
- 1991
(Show Context)
Citation Context ...cribed in [2]. Mesh "edges" provide a rich potential source of input-output bandwidth for visualization, or disk farms. methods for communication primitives and collections of communicating =-=processes [14, 18]-=-. A lot more can be said about the communication issues that the Toolbox addresses and supports, and that underly grid-based communication described here, as a particular example; this discussion appe... |

5 |
de Velde. Experiments with Multicomputer LU-Decomposition. Concurrency: Practice and Experience
- van
- 1990
(Show Context)
Citation Context ...arameters) to incorporate practically important degrees of freedom. These generalized distribution functions yield optimal static load balance as do the ungeneralized functions described here (and in =-=[25]-=-) for unit block size, but differ in coefficient placement. This distinction is necessary for efficient implementations. 4.3.1 Conventional Functions Definition 4.4 (Linear) The conventional linear, l... |

3 | factorization of sparse, unsymmetric Jacobian matrices on multicomputers: Experience, strategies, performance - LU - 1990 |

3 |
Concurrent DASSL applied to dynamic distillation column simulation
- Skjellum, Morari
- 1990
(Show Context)
Citation Context ...l work) to address more complex flowsheeting problems in chemical engineering; fourth, ideas for how to achieve higher performance with Cdyn, using iterative methods for the underlying linear algebra =-=[16, 15, 17, 14]-=-. Of these, the chief emphasis in the present paper is the reusable collection of parallel libraries comprising the Toolbox. Concurrent algorithms for the solution of dense and sparse linear systems, ... |

1 |
Waveform Relaxation for Concurrent Dynamic Simulation of Distillation Columns
- Skjellum, Morari, et al.
- 1988
(Show Context)
Citation Context ...lgorithms for engineering applications on message-passing concurrent computers (multicomputers). At Caltech, the first author studied the concurrent dynamic simulation of distillation column networks =-=[19, 21, 20, 14]-=-. This research was accomplished with attention to portability, high performance and reusability of the underlying algorithms. Emerging from this work are several key results: first, a methodology for... |

1 |
Highly Concurrent Dynamic Simulation
- Skjellum, Morari, et al.
- 1988
(Show Context)
Citation Context ...lgorithms for engineering applications on message-passing concurrent computers (multicomputers). At Caltech, the first author studied the concurrent dynamic simulation of distillation column networks =-=[19, 21, 20, 14]-=-. This research was accomplished with attention to portability, high performance and reusability of the underlying algorithms. Emerging from this work are several key results: first, a methodology for... |

1 |
Application of Multicomputers to Large-Scale Dynamic Simulation
- Skjellum, Peterson, et al.
- 1989
(Show Context)
Citation Context |

1 |
de Velde. Implementation of Linear Algebra Operations on Multicomputers
- Van
- 1988
(Show Context)
Citation Context ...lization of x to zero at the outset allows for the negligible "bookkeeping" inherent in this conversion procedure. 5.4 Inner Products First we define two conventional inner products per, for=-= example, [23, 24]-=-. Then we consider a new kernel, the skew inner product, and a communication-free kernel, skew outer product. Definition 5.5 (Inner Product) The simplest inner product x T y of two concurrent vectorss... |

1 |
Parallel Waveform Relaxation Methods for Solving Parabolic Partial Differential Equations
- Vandewalle
- 1990
(Show Context)
Citation Context ...enerates a better sequential algorithm. In this case, we must relabel our diagram so that, again, the "best" sequential algorithm has the lowest time for the single-processor limit (see, for=-= example, [26]). Otherwi-=-se, we are artificially inflating our results. The "Concurrency Diagram" is a powerful means of expression of concurrency ideas, including the ability to compare algorithms run on different ... |