Computers for correcting structural errors in spectral information
β Scribed by O. D. Gorokhov; V. N. Nikitin
- Publisher
- Springer US
- Year
- 1975
- Tongue
- English
- Weight
- 338 KB
- Volume
- 23
- Category
- Article
- ISSN
- 0021-9037
No coin nor oath required. For personal study only.
π SIMILAR VOLUMES
## Abstract This paper proposes an algorithm to accurately compute the sum of floatingβpoint numbers on parallel computers. This algorithm is an extension of the wellβknown recursive doubling technique which computes the sum of __n__ floatingβpoint number in log~2~__n__ parallel steps. The time com
TlLe errors ipa structural com@ations are a,ssessed and bounds aye established on their magnitude. The matrix decomposition approuch is shown to be most useful in error analysis. Both positive definite and positive semi-dejkite matrices are treated. T?ie &scretefLnite-digit representation qf Ilumber