Abstract | ||
---|---|---|
We show that if the size of the tridiagonal matrix in any given iteration is n, then the parallel QR algorithm requires O(log2n) steps with O(n) processors per iteration and no square roots. This results in a speedup of O(n/log2n) over the sequential algorithm with an efficiency of O(1/log2n). We also give an error analysis of the parallel triangular system solvers used in each iteration. |
Year | DOI | Venue |
---|---|---|
1977 | 10.1109/TC.1977.5009293 | IEEE Trans. Computers |
Keywords | Field | DocType |
recurrence relations,tridiagonal matrix,algorithm design and analysis,data mining,recurrence relation,symmetric matrices,probability density function | Tridiagonal matrix,Discrete mathematics,Matrix (mathematics),Symmetric matrix,Sequential algorithm,Mathematics,Tridiagonal matrix algorithm,Power iteration,QR algorithm,Speedup | Journal |
Volume | Issue | ISSN |
26 | 2 | 0018-9340 |
Citations | PageRank | References |
16 | 18.84 | 0 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ahmed H. Sameh | 1 | 297 | 139.93 |
David J. Kuck | 2 | 625 | 513.07 |