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

Re: mathematica & fortran ??

• To: mathgroup at smc.vnet.net
• Subject: [mg71426] Re: mathematica & fortran ??
• From: "Beliavsky" <beliavsky at aol.com>
• Date: Sat, 18 Nov 2006 04:40:57 -0500 (EST)
• References: <ejjvg3$l3f$1@smc.vnet.net>

fish wrote:
> I have been using the software of Mathematica for 2 years. It is a
> good software.
>
>    Those days, I meet a difficulty. I compile a program, which is
> written in Fortran language. In it, there is a Bessel function. I want
> to get the high precision value of the function. Unfortunately, I
> failed.
>
>    As we know, the software of Mathematica supplies the Bessel function
> with high precision. So I want to ask whether we can run the two
> softwares, Mathematica and Fortran language, simultaneously? Then I can
> directly use the Bessel function in Mathematica for my Fortran program.

A Fortran solution could be to find a compiler that supports quadruple
precision (16 byte floating point variables) or 10-byte reals. Then you
could either change the code to use quadruple precision or (depending
on the compiler) invoke a compiler switch that automatically promotes
double precision variables to quadruple precision. The Lahey/Fujitsu
compiler has the -quad option. If you want to try this route, the best
newsgroups for follow-up questions would be comp.lang.fortran and/or
sci.math.num-analysis.