HOWTO: Convert 10-Byte Long Doubles to 8-Byte Doubles

• The Microsoft C/C++ Compiler (CL.EXE) included with: - Microsoft Visual C++, 32-bit Edition, versions 1.0, 2.0, 2.1, 4.0, 5.0

       on the following platform:
       - x86
  

SUMMARY

With the 16-bit Microsoft C/C++ compilers, long doubles are stored as 80- bit (10-byte) data types. Under Windows NT, in order to be compatible with other non-Intel floating point implementations, the 80-bit long double format is aliased to the 64-bit (8-byte) double format.

This means that 32-bit programs may not be able read back data files written by 16-bit programs because the long double formats are incompatible.

On Intel platforms, the only workaround is to let the floating point processor handle the conversion from 80-bit to 64-bit doubles. Afterwards, the data can be stored back into a 64-bit double for use under Win32.

The sample code below illustrates how you could use floating point instructions in inline assembly to convert from a 10-byte double in a data file to an 8-byte double.

Sample Code

   /* Compile options needed: none
   */

   #include <stdio.h>

   void main(void)
   {
      FILE *inFile;
      char buffer[10];
      long double Newdbl;

      inFile = fopen("data","rb");
      fread(buffer, 10, 1, inFile);      // reads in 10-byte long double
      fclose(inFile);

      // This moves the contents of the buffer into the floating point
      // register, which then then takes care of the automatic convertion
      // back to a 8-byte long double

      _asm {
         fld TBYTE PTR buffer;
         fstp Newdbl;
      }
   }