Loading...
Please wait, while we are loading the content...
Similar Documents
Parallel random number generation and high performance fortran (1993).
| Content Provider | CiteSeerX |
|---|---|
| Author | Hennecke, Michael |
| Abstract | Michael Hennecke, Parallel Random Number Generation and High Performance Fortran, Submitted to Parallel Computing. Parallel random number generation is discussed with the proviso of generating fully reproducible random number streams on an arbitrary number of processors. The demand that data distributions onto processors should be transparent to the user is taken into account by utilizing High Performance Fortran specifications to allow the user to explicitly specify a data distribution suitable for his or her application. The random number generator "inherits" the distribution and automatically chooses an algorithm that works optimally for that distribution of the random numbers. These concepts have been put to work by implementing a parallel version of the random number generation algorithm used in the NAG Fortran library on a Kendall Square KSR1 parallel computer, for which detailed performance measurements are presented. Keywords. Random number generation; High Performance Fortr... |
| File Format | |
| Publisher Date | 1993-01-01 |
| Access Restriction | Open |
| Subject Keyword | Parallel Random Number Generation High Performance Fortran Detailed Performance Measurement Parallel Computing Data Distribution Suitable Arbitrary Number Nag Fortran Library Reproducible Random Number Stream Random Number High Performance Fortr High Performance Fortran Specification Parallel Version Kendall Square Ksr1 Random Number Generation Algorithm Random Number Generator Random Number Generation Data Distribution |
| Content Type | Text |
