Results Differ for Alternate & Coprocessor/Emulator Math

MS-DOS                      | OS/2            | WINDOWS

The information in this article applies to:

• The C Run-time (CRT) included with:

      - Microsoft C for MS-DOS, versions 5.1, 6.0, 6.0a, and 6.0ax
      - Microsoft C for OS/2, versions 5.1, 6.0, and 6.0a
      - Microsoft C/C++ for MS-DOS, version 7.0
      - Microsoft Visual C++ for Windows, version 1.0
  

SUMMARY

The coprocessor/emulator math library and the alternate math library may produce different results for floating-point calculations.

MORE INFORMATION

When an application uses the coprocessor to perform floating-point arithmetic, it pushes all values onto the coprocessor stack. The stack holds only 10-byte real values. Therefore, any 4-byte or 8-byte real (double) values are expanded to 10-byte reals before any operations are performed. The emulator library performs the same expansion. The results are converted back to a 4-byte or 8-byte real value as appropriate.

The alternate math library performs calculations directly using 4-byte or 8-byte real values which can result in less calculation precision than that available with the coprocessor/emulator math library.