Parallel singular value decomposition with cyclic storing
✍ Scribed by Norbert Köckler; Matthias Simon
- Publisher
- Elsevier Science
- Year
- 1991
- Tongue
- English
- Weight
- 343 KB
- Volume
- 17
- Category
- Article
- ISSN
- 0167-8191
No coin nor oath required. For personal study only.
✦ Synopsis
K~Sckler, N. and M. Simon, Parallel singular value decomposition with cyclic storing, Parallel Computing 19 (1991) 39-47.
The singular value decomposition A = UY'V r is implemented on a processor network. The Hestenes and the Golub-Reinsch-algorithms axe compared. If the columns of A are distributed to the processors in cyclic order (instead of blockwise storing), the Golub-Reinsch-algorithm (GRA) becomes much more efficient and much faster than the Hestenes algorithm. The GRA is almost independent of the structure of the processor net. With growing matrix orders load balancing and speedup of the cyclic parallel GRA is almost optimal.
📜 SIMILAR VOLUMES
A number of modified two-dimensional NMR pulse sequences FIDs in the indirectly detected (t 1 ) domain after the Fourier have been developed to suppress the diagonal peaks which may transformation of the directly detected (t 2 ) domain. These introduce severe spectral distortions, such as t 1 -ridge
## Abstract A new MRI spatial encoding method based upon the singular value decomposition (SVD) and using spatially selective RF excitation is described. This encoding technique is particularly; applicable to dynamic adaptive MRI, because it provides a near minimal set of spatial encoding profiles