11 months ago
Improved the documentation.
| lib/impl_SD.mli | file | annotate | diff | revisions |
1.1 --- a/lib/impl_SD.mli 1.2 +++ b/lib/impl_SD.mli 1.3 @@ -675,7 +675,10 @@ 1.4 ?ac : int -> 1.5 mat 1.6 -> vec 1.7 -(** [syev ?n ?vectors ?up ?ofswork ?work ?ofsw ?w ?ar ?ac a] 1.8 +(** [syev ?n ?vectors ?up ?ofswork ?work ?ofsw ?w ?ar ?ac a] computes 1.9 + all eigenvalues and, optionally, eigenvectors of the real symmetric 1.10 + matrix [a]. 1.11 + 1.12 @return the vector [w] of eigenvalues in ascending order. 1.13 @raise Failure if the function fails to converge. 1.14 @param n default = available number of columns of matrix [a] 1.15 @@ -759,6 +762,10 @@ 1.16 mat 1.17 -> vec 1.18 (** [syevd ?n ?vectors ?up ?ofswork ?work ?iwork ?ofsw ?w ?ar ?ac a] 1.19 + computes all eigenvalues and, optionally, eigenvectors of the real 1.20 + symmetric matrix [a]. If eigenvectors are desired, it uses a 1.21 + divide and conquer algorithm. 1.22 + 1.23 @return the vector [w] of eigenvalues in ascending order. 1.24 @raise Failure if the function fails to converge. 1.25 @param n default = available number of columns of matrix [a] 1.26 @@ -973,7 +980,7 @@ 1.27 (** [sbgv ?n ?ka ?kb ?zr ?zc ?z ?up ?work ?ofsw ?w ?ar ?ac a ?br ?bc b] 1.28 computes all the eigenvalues, and optionally, the eigenvectors of a 1.29 real generalized symmetric-definite banded eigenproblem, of the 1.30 - form A*x=(lambda)*B*x. Here [a] and [b] are assumed to be 1.31 + form [a*x=(lambda)*b*x]. Here [a] and [b] are assumed to be 1.32 symmetric and banded, and [b] is also positive definite. 1.33 1.34 @return the vector [w] of eigenvalues in ascending order.