[Octave-bug-tracker] [bug #51950] sparse qr may return a malformed sparse matrix R with out of bounds entries

[Rik]

Follow-up Comment #15, bug #51950 (project octave):

Good analysis.  One question, is it always the case that the inserted rows can
safely be deleted?  In the example we are using, the R matrix has three fully
empty rows between the last row with an entry at (2,2) and the next entry at
(6,6).  Collapsing null rows is a fairly easy operation and if the entry at
(6,6) were moved to (3,6) everything would work out correctly.

Second, Tim Davis, author of SuiteSparse has moved from the U. of Florida to
Texas A&M.  The overview page for SuiteSparse is now at
http://faculty.cse.tamu.edu/davis/suitesparse.html.  According to this page,
Matlab *has* moved QR to rely on the SuiteSparse library.  Given that Octave
already uses SuiteSparse in some capacity, maybe we should just switch to a
different library.  Is there a comparison table between CSparse, CXSparse, and
SuiteSparse in terms of things like memory usage, efficiency, speed, etc.?

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?51950>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/