Adding Harwell-Boeing support in PETSc?

[Matthew Knepley]

Richard Tran Mills <rmills at ornl.gov> writes:

> Hi Folks,
>
> It seems that someone always wants me to either give them a matrix in Harwell-Boeing format, or to try experimenting
> with a matrix that they give me in Harwell-Boeing format.  I would prefer that the format just go away, but since this
> doesn't seem to be happening, I would like to have some PETSc support mechanisms for reading and writing the beasts.
> I'd like to have something like a PETSC_VIEWER_HB for writing them out, and I'd like to have MatLoad() be able to use
> the viewer to load the matrices as well.
>
> I thought I might look into adding something like this, but before I waste time with it, I thought I would ask to folks
> if either a) they think it is a bad idea or b) someone has already got the code for this.

  Well, Mat example 81 saves a matrix in HB format. The other way shuold not be too hard if you do not make it too
general. A viewer would be nice if you feel like doing it.

        Matt
-- 
"Failure has a thousand explanations. Success doesn't need one" -- Sir Alec Guiness