MathGroup Archive 2005

[Date Index] [Thread Index] [Author Index]

Search the Archive

Mathematica Performance

  • To: mathgroup at smc.vnet.net
  • Subject: [mg62413] Mathematica Performance
  • From: Bill Rowe <readnewsciv at earthlink.net>
  • Date: Wed, 23 Nov 2005 01:13:11 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

On 11/22/05 at 4:42 AM, hattons at globalsymmetry.com (Steven T.
Hatton) wrote:

>For purposes of number crunching Mathematica is quite often the
>wrong tool for the job.  If you are doing things which involve very
>intensive mumeric calculation which are likely to be repeated in
>form, you are probably much better off translating your work to
>something like C++, or FORTRAN. 

It is certainly true a well written C++ or FORTRAN code compiled correctly will result in a program that executes much faster than the corresponding Mathematica code for a non-trivial algorithm. But I disagree this implies a user is better off translating his work to C++ or FORTRAN than using Mathematica.

Program execution speed determines only a portion of the time the user needs to wait until he gets his answer. For most problems, it is likely the time needed to determine what algorithm is appropriate, write the code and debug it is far greater than the time needed for the resulting program to run.

And the effort involved in translating well written, efficient Mathematica code to C++ or FORTRAM will be far from trivial. There is no equivalent in either C++ or FORTRAN to Mathematicas Apply, Map functions, replacement rules etc. Translating these functions to C++ or FORTRAN not only takes significant time but also greatly increases the likelihood of bugs. This is particularly true since someone skilled at creating Mathematica code is unlikely to be equally skilled at writing C++ or FORTRAN code.

Skill in a particular language comes about from using that language extensively. That skill tends to atrophy when the language is not used. And clearly, someone spending most of his time working with Mathematica is likely far more skilled with Mathematica than either C++ or FORTRAN.
--
To reply via email subtract one hundred and four


  • Prev by Date: Re: Boolean Integral
  • Next by Date: Re: How to View Mathematica and Hardcopy Books
  • Previous by thread: Re: Re: Time needed for calculation
  • Next by thread: MATHwire: Information for Mathematica Users