Computational and Theoretical Techniques for Materials Science (1995) / Chapter Skim
Currently Skimming:
Chapter 3--Computational Perspectives for the Remainder of the Decade
Chapter 3--Computational Perspectives for the Remainder of the Decade
Pages 34-39
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 34... ...
Quantum molecular (Iynamical simulations, such as those employing the Car-Parrinello method, have been successfully parallelized, despite the fact that these methods make extensive use of the communication-demanding fast Fourier transform. These applications share with the larger class of nonlinear optimization methods, such as those based on functional gradients, the important distinction between the parallelization of inner anti outer loops.
|
From page 35... ...
The importance of communication requirements stems from the fact that the time required to retrieve data from the memory of another processor is very much longer than from local memory. Some algorithms, such as the fast Fourier transform, require extensive interprocessor communication.
|
From page 36... ...
Computational materials science is one of the fields already benefiting from this increase in computational power available at national sites. A summary of some of the current ant!
|
From page 37... ...
It is worth restating that this big increase in computational power is due not only to using more processors but also to an order-ofmagnitude increase in the speed of individual processors. It is important to keep this in mind since a large number of problems of interest in computational materials science (e.g., many using molecular dynamics)
|
From page 38... ...
The most sophisticated debuggers are integrated graphical programming environments within which users can develop, edit, execute, debug, and profile the performance of programs written either in message passing or for global parallel execution. Features such as the visual display of timing, bottlenecks, and so on greatly facilitate the optimization of parallel code.
|
From page 39... ...
Such reduced data sets will allow remote users to benefit from the power and versatility of MPPs and parallel visualization over the network. The arrival of very fast network connections will alleviate many of the connection problems faced by remote users and the data transfer problems among computers.
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.