| Matrix properties | |
| number of rows | 1,824 |
| number of columns | 1,824 |
| nonzeros | 39,208 |
| structural full rank? | yes |
| structural rank | 1,824 |
| # of blocks from dmperm | 1 |
| # strongly connected comp. | 1 |
| entries not in dmperm blocks | 0 |
| explicit zero entries | 0 |
| nonzero pattern symmetry | symmetric |
| numeric value symmetry | symmetric |
| type | real |
| structure | symmetric |
| Cholesky candidate? | yes |
| positive definite? | no |
| author | R. Grimes |
| editor | T. Davis |
| date | 1995 |
| kind | duplicate structural problem |
| 2D/3D problem? | yes |
| Additional fields | size and type |
| b | full 1824-by-1 |
Notes:
Let A1=Nasa/nasa1824 and A2=Boeing/nasa1824. A1 and A2 have the same nonzero pattern. A1 and A2 differ in value in only 386 entries out of 39208, and only in 21 columns of the lower triangular part; tril(A(196:321,196:216)) and the same rows of the upper triangular part. The magnitudes of the entries in A2 in this region of the matrix are all tiny, and have only 9 digits if printed in base-10 (unlike the other entries, which have full precision). I suspect A2 (Boeing/nasa1824) is a corrupted version of A1 (Nasa/nasa1824).
| Ordering statistics: | AMD | METIS |
| nnz(chol(P*(A+A'+s*I)*P')) | 69,309 | 79,854 |
| Cholesky flop count | 4.6e+06 | 5.8e+06 |
| nnz(L+U), no partial pivoting | 136,794 | 157,884 |
| nnz(V) for QR, upper bound nnz(L) for LU | 157,808 | 138,536 |
| nnz(R) for QR, upper bound nnz(U) for LU | 289,058 | 259,834 |
Maintained by Tim Davis, last updated 04-May-2008.
Matrix pictures by cspy, a MATLAB function in the CSparse package.
Matrix graphs by Yifan Hu, AT&T Labs Visualization Group.