Abstract
In this short note we describe some of the problems we have encountered on implementing the R-Matrix codes of Berrington et al. on the CRAY-1. In particular we consider the calculation of R-matrix angular integrals which as currently implemented are not vectorizable. Since this section of the code is sandwiched between highly vectorizable packages a bottleneck occurs. An alternative algorithm to circumvent this impasse is suggested.
Original language | English |
---|---|
Pages (from-to) | 419-421 |
Number of pages | 3 |
Journal | Computer Physics Communications |
Volume | 26 |
Issue number | 3-4 |
DOIs | |
Publication status | Published - Jun 1982 |
ASJC Scopus subject areas
- Hardware and Architecture
- General Physics and Astronomy