Re: Looking for gen symm sparse eigensolver in C or C++

From: Madhusudan Singh (spammers-go-here_at_spam.invalid)
Date: 03/28/05

  • Next message: glen herrmannsfeldt: "Re: Namelist Internal Read" 
    Date: Sun, 27 Mar 2005 22:30:34 -0500

    carlos@colorado.edu wrote:

    > Victor Eijkhout wrote:
    >> <carlos@colorado.edu> wrote:
    >>
    >> Arpack.
    >>
    >> > Fortran implementations excluded.
    >>
    >> I have no idea what language Arpack is written in. Why on earth does
    > it
    >> matter? Reading Fortran is trivial, and knowing how to link it into a
    > C
    >> program is something every student around here knows how to do.
    >>
    >> V.
    >> --
    >> email: lastname at cs utk edu
    >> homepage: www cs utk edu tilde lastname
    >
    > It has to be implemented by a student who is proficient in C and C++.
    >
    > Few American undergraduate or graduate students know about Fortran, or
    > care. The teaching of Fortran stopped in our College of Engineering in
    > 1997. (It is still taught in Chemistry and Physics - they are in the
    > College of Arts & Science - because of large amounts of legacy code)
    > Reason: industry reps in the College EAB convinced faculty that it is a
    > waste of time - they say that legacy code still in use can be
    > maintained by old timers. Proficiency in C, Java, matlab, idl, labview
    > and excel were judged more career helpful.
    >
    > I disagreed for sentimental reasons (my first programming languages
    > were Fortran II and MAP on a 7094) but there is no fighting reality.

    In that case, the members of that "brain trust" can live with the reality as
    well. There is nothing you need in scientific computation that C++ has and
    Fortran 95 doesn't. Including pointers, recursion, operator overloading,
    OOP, etc.

    You do not need any sentimental reasons to stick to Fortran. The code is far
    easier to understand, less likely to get sphagettized and older fortran
    code works seamlessly with the newer versions. Only a top class moron would
    want to invest the time in re-inventing the wheel and junking code that has
    been battle tested over decades.

    Returning to your OP, look at arpack++. I have heard that it is written in
    C++, but the I do not know how well tested it is. Caveat Emptor.

  • Next message: glen herrmannsfeldt: "Re: Namelist Internal Read"

    Relevant Pages

    • Re: calling ARPACK libraries from C++
      ... > I am trying to use ARPACK and BLAS libraries by calling them from my C++ ... Are there any online resources on how to do this? ... I learned FORTRAN a quite a few years ago and haven't practiced it ...
      (comp.lang.fortran)
    • Re: Looking for gen symm sparse eigensolver in C or C++
      ... >> I have no idea what language Arpack is written in. ... Reading Fortran is trivial, and knowing how to link it into a ... You do not need any sentimental reasons to stick to Fortran. ...
      (sci.math.num-analysis)
    • Re: calling ARPACK libraries from C++
      ... Are there any online resources on how to do this? ... If I can get example programs about the FORTRAN~C++ interface regarding ARPACK, ... I learned FORTRAN a quite a few years ago and haven't practiced it ... ARPACK library directly? ...
      (comp.lang.fortran)
    • calling ARPACK libraries from C++
      ... Are there any online resources on how to do this? ... If I can get example programs about the FORTRAN~C++ interface regarding ARPACK, ... I learned FORTRAN a quite a few years ago and haven't practiced it since. ...
      (comp.lang.fortran)
    • Re: Looking for gen symm sparse eigensolver in C or C++
      ... >> Fortran implementations excluded. ... > I have no idea what language Arpack is written in. ... It has to be implemented by a student who is proficient in C and C++. ... College of Arts & Science - because of large amounts of legacy code) ...
      (sci.math.num-analysis)