-
Notifications
You must be signed in to change notification settings - Fork 475
Closed
Labels
Description
I am not aware of a text that defines a minor to be a matrix. Hence I was bit surprised to read
the leading minor of order i is not positive definite
here:
Lines 87 to 92 in cea4a63
*> INFO is INTEGER | |
*> = 0: successful exit | |
*> < 0: if INFO = -i, the i-th argument had an illegal value | |
*> > 0: if INFO = i, the leading minor of order i is not | |
*> positive definite, and the factorization could not be | |
*> completed. |
Actually, grep -n SRC/*.f -e "minor"
reveals many more instances ...
AFAIK, a minor is a determinant and not a matrix (see, e.g., Aitken, 1944, p. 39; Horn and Johnson, 2013, p. 17; ...), and it is therefore nonsense to consider whether a minor is positive definite. Should the text be globally amended to
the leading principal minor of order i is not positive
or
the leading principal submatrix of order i is not positive definite
depending on context? It would be the former in the case of DPOTRF
, at least according to Sylvester's criterion ...
- I've included a minimal example to reproduce the issue
- I'd be willing to make a PR to solve this issue