NAG Technical Report 3/2009

Calling NAG Library Routines from Octave

3.2. Example 2

Scaled complex complement of error function, exp(-z2)erfc(-iz) - nag_complex_erfc (s15ddc) routine

Contents

1. Function prototype from the NAG C Library Manual
2. C++ function
3. Compiling into Oct-File
4. Calling the function

1. Function prototype from the NAG C Library Manual

   According to the C Library Manual, the prototype for function s15ddc looks like this:

     #include <nag.h>
     #include <nags.h>
   
     Complex nag_complex_erfc(Complex z, NagError *fail);
   

   The function takes two arguments. The first one is of type Complex and is the argument z of the function. The second argument to s15ddc is the error handling argument fail which is of type NagError. See the NAG C Library Manual for more information about the NagError type. To keep things simple we are not going to try to pass the contents of the NagError structure back to Octave. Instead, we will return only integer value of fail.code.

2. C++ function

   Here is the source code of our C++ function nag_cmplx_erfc.cc:

   #undef Complex
   #define Complex OctComplex
   #include <octave/oct.h>
   #undef Complex
   #define Complex NagComplex
   #define MatrixType NagMatrixType
   #include <nag.h>
   #include <nags.h>
   
   DEFUN_DLD (nag_cmplx_erfc, args, ,
     "Calls nag_complex_erfc (s15ddc), which computes values \n\
   of the function w(z)=e^(-z^2)erfc-iz, for complex z.\n")
   {
    // Variable to store function output values
    octave_value_list retval;
    // Retrieve input arguments from args
    OctComplex z = args(0).complex_value();
    // Local variables
    NagError fail;
    INIT_FAIL(fail);
    OctComplex w;
    NagComplex z_nag, w_nag;
    // Copy the input OctComplex variable into NagComplex one
    z_nag.re = z.real();
    z_nag.im = z.imag();
         
     // Call NAG routine like in C
     w_nag = nag_complex_erfc(z_nag,&fail);
       
     // Copy the output NagComplex variable into OctComplex one
     w.real() = w_nag.re;
     w.imag() = w_nag.im;
     // Assign output arguments to retval
     retval(0) = w;
     retval(1) = fail.code;
        
     return retval;
   }
   

   Points to note about this code:

   • For clarity, the input arguments type checking and error message printing have been omitted in the article text. However, they have been added to the file you can download from this site.

   • Complex and MatrixType types are defined in Octave and NAG header files. In this example we are using Complex type, so we need to rename both of its definitions. We overwrite it with OctComplex, before including oct.h, and with NagComplex, before including nag.h. We are not using MatrixType, so it can be renamed in one place only, like in the previous example.

   • NAG and Octave Complex types are defined differently, so we cannot pass an Octave complex variable directly to a NAG routine. We create a local variable of type NagComplex and make a copy of the Octave variable. We do the opposite with the complex type result from the NAG routine.

   • The third DEFUN_DLD argument (nargout) is not used, so it is omitted in order to avoid the warning from gcc about an unused function parameter.

3. Compiling into Oct-File

   To compile the C++ function into oct-files, we use the mkoctfile script supplied with Octave:

     % mkoctfile nag_cmplx_erfc.cc -L/opt/NAG/cll6a08dg/lib -lnagc_nag -I/opt/NAG/cll6a08dg/include
   

   where:
   
   • the -L switch tells the C compiler where to look for NAG C Library installed on your system;
     
   • -lnagc_nag is the name of the NAG C Library;
     
   • the -I switch should be followed by the location of NAG C Library header files.

4. Calling the function

   Assuming that all has gone well, we can call the function as if it was part of Octave itself, i.e. either from the Octave command line or from within an Octave program. An example call may be:

   octave:1> z=complex(-3.01,0.75)
   z = -3.01000 + 0.75000i
   octave:2> [w,ifail]=nag_cmplx_erfc(z)
   w =  0.052156 - 0.183776i
   ifail = 0
   

   (If you get an error message saying that a library cannot be located, see the tip given in Example 1).