You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In order to implement the psd projection of a matrix in packed storage format, i.e. upper triangular part of the matrix stacked in a vector, we should make use of the corresponding LAPACK function. Otherwise, we will have to "unpack" the matrix every time we want to project, check infeasibility, etc.
http://www.icl.utk.edu/~mgates3/docs/lapack.html
gives a good overview on the available LAPACK functions for eigenvalue computation. It seems like the relevant functions are spev, spevx and spevd.
Unfortunately, the julia LAPACK package doesn't provide wrappers for any of these functions. However, the wrapper for e.g. syevr doesn't seem to be too complicated:
In order to implement the psd projection of a matrix in packed storage format, i.e. upper triangular part of the matrix stacked in a vector, we should make use of the corresponding LAPACK function. Otherwise, we will have to "unpack" the matrix every time we want to project, check infeasibility, etc.
http://www.icl.utk.edu/~mgates3/docs/lapack.html
gives a good overview on the available LAPACK functions for eigenvalue computation. It seems like the relevant functions are
spev
,spevx
andspevd
.Unfortunately, the julia LAPACK package doesn't provide wrappers for any of these functions. However, the wrapper for e.g.
syevr
doesn't seem to be too complicated:LAPACK.jl Documentation - Line 4949
So I think we should write our own wrapper for one of the above functions.
The text was updated successfully, but these errors were encountered: