[Octave-bug-tracker] [bug #46170] David Bateman: sparse matrix * single

[Ceral Paquet]

URL:
  <http://savannah.gnu.org/bugs/?46170>

                 Summary: David Bateman: sparse matrix * single
                 Project: GNU Octave
            Submitted by: octavebugs
            Submitted on: Sat 10 Oct 2015 08:31:58 AM GMT
                Category: None
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: None
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: 4.0.0
        Operating System: GNU/Linux

    _______________________________________________________

Details:

I am interested in contributing functionality to the Octave project. I saw
your name in the comments for sparse functions and searched online and found
your pdf article on the Octave implementation.

Of most interest to me would be to allow the multiplication of a
sparse matrix with a single precision vector (or matrix). to return a full
single precision vector (or full matrix). The reason is that sparse requires
everything in double precision and so all my data (multi Gb in single
precision) have to be cast up to doubles just to do a multiplication.

This does extend the functionality as compared to Matlab but doesn't break
compatibility - I think it is much like how Octave extends useful things like
# for comments and += operators. Also, as it happens, how Octave's sparse()
accepts non-double arguments whereas Matlab's gives an error.

I've been staring at the Sparse code in liboctave/array for a few days and
tried a few things but I am lost! I wonder if you could give me a short intro
to the sparse classes:

boolSparse.cc  dSparse.cc  MSparse-C.cc  Sparse-b.cc  Sparse-C.cc
CSparse.cc     MSparse.cc  MSparse-d.cc  Sparse.cc    Sparse-d.cc
boolSparse.h  CSparse.h  dSparse.h  MSparse-defs.h  MSparse.h  Sparse.h

I see operator definitions for * in different files and some FORTRAN
interfaces but it's hard to understand what does what. I'd really appreciate
your help.

Thanks!





    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?46170>

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