[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-patch-tracker] [patch #8542] Handle sparse matrices in cholinser
From: |
Daniel Kraft |
Subject: |
[Octave-patch-tracker] [patch #8542] Handle sparse matrices in cholinsert |
Date: |
Wed, 24 Sep 2014 05:15:25 +0000 |
User-agent: |
Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:32.0) Gecko/20100101 Firefox/32.0 |
URL:
<http://savannah.gnu.org/patch/?8542>
Summary: Handle sparse matrices in cholinsert
Project: GNU Octave
Submitted by: domob
Submitted on: Wed 24 Sep 2014 07:15:24 AM CEST
Category: None
Priority: 5 - Normal
Status: None
Privacy: Public
Assigned to: None
Originator Email:
Open/Closed: Open
Discussion Lock: Any
_______________________________________________________
Details:
Octave doesn't currently handle sparse matrices in cholinsert (or
cholupdate/choldelete). The functionality is supported by CHOLMOD, though.
The attached patches implement cholinsert for sparse matrices. It will be
easy to extend them to cholupdate and choldelete, also. In addition to the
usage for dense matrices, it is also possible to "bulk insert" multiple rows
at once--this is more efficient than calling cholinsert repeatedly, since the
conversion to/from cholmod_factor can be skipped.
According to discussions with the CHOLMOD maintainer, the current
implementation has problems if exact zeros were dropped in the Cholesky
factor. It seems to work fine for my own purposes so far, but it may be
necessary to fix that. A possible idea is to do a symbolic factorisation of
the matrix (which is fast) and then use the resulting sparsity pattern to fill
in exact zeros in L. The sparsity pattern could even be returned as an
optional output of chol. Users of cholinsert could then just keep it
alongside of L and pass it to the sparse version of cholinsert.
_______________________________________________________
File Attachments:
-------------------------------------------------------
Date: Wed 24 Sep 2014 07:15:24 AM CEST Name: patch_1.diff Size: 10kB By:
domob
First version of a patch set.
<http://savannah.gnu.org/patch/download.php?file_id=32175>
-------------------------------------------------------
Date: Wed 24 Sep 2014 07:15:24 AM CEST Name: patch_2.diff Size: 11kB By:
domob
First version of a patch set.
<http://savannah.gnu.org/patch/download.php?file_id=32176>
-------------------------------------------------------
Date: Wed 24 Sep 2014 07:15:24 AM CEST Name: patch_3.diff Size: 17kB By:
domob
First version of a patch set.
<http://savannah.gnu.org/patch/download.php?file_id=32177>
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/patch/?8542>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Octave-patch-tracker] [patch #8542] Handle sparse matrices in cholinsert,
Daniel Kraft <=