What are the data dependency issues

Parallelization of the algorithms

Parallel Linear Algebra pp 31-47 | Cite as

Summary

Only the upper levels are for the user of parallel computer structures (Vector processing and concurrency See Chapter 1.) of parallelism is important because he can use it with his programs or has a direct influence on the efficiency of the implementation. In theory, these levels should also be user-transparent. In practice, however, it has been shown that the automatic parallelization of solution processes is still in its infancy, especially when one considers concurrency. The reasons for this are briefly discussed in the first part of this chapter, which deals with data dependency analysis. The parallelization of algorithms is a very young discipline, which is why many of the terms used have not yet been standardized. In addition, most of the terms come from English and are not always easy to translate. A brief summary of the terms used in this book is intended to avoid confusion:

This is a preview of subscription content, log in to check access.

Preview

Unable to display preview. Download preview PDF.

Copyright information

© Deutscher Universitäts-Verlag GmbH, Wiesbaden 1992

Authors and Affiliations

There are no affiliations available