-
-
Notifications
You must be signed in to change notification settings - Fork 5.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
WIP: ENH: linalg: add wrapper for ?pbsvx
#11985
Conversation
|
Ah nevermind it's looking for an identical scaling on both sides. Then yes |
828ab58
to
c82ebf4
Compare
@mdhaber @Kai-Striega I assume that this is abandoned as the corresponding issue was closed as not planned. Please reopen if I'm mistaken |
?pbsvx
@lucascolley Yes, I think this can be closed. |
Reference issue
Closes issue #11616
What does this implement/fix?
This PR adds a python wrapper for the LAPACK functions {d,s,c,z}pbsvx. These functions solve the linear system of equations Ax = b where A is a Symmetric/Hermitian matrix.
Additional information
?pbsvx comes from a more recent version of lapack than our current minimum. It should therefore not yet be merged.
This PR is missing tests with
equed='Y'
. This test requiresA
to be multiplied by some scaling factors
i.e.A = diag(s) @ A @ diag(s)
. The problem I'm having is how do I best generate s? If I were to use a random s, the resultant matrix is (almost certainly) no longer hermitian alternatively I could use ?geequb this returns row, column factors, in that case, how should I convert the row, col factors tos
?