Q36069: Different Results between Alternate and Coprocessor/Emulator Math

Article: Q36069
Product(s): See article
Version(s): 3.00 4.00 5.00 5.10 | 5.10
Operating System(s): MS-DOS | OS/2
Keyword(s): ENDUSER | s_pascal h_fortran h_masm | mspl13_c
Last Modified: 12-OCT-1988

There may be differences between the floating point values you get
when using coprocessor/emulator math and the alternate math packages.

This behavior applies to C Versions 3.00, 4.00, 5.00, and 5.01, Pascal
Versions 3.1x, 3.20, 3.30, 3.31, 3.32, and 4.00, FORTRAN Versions
3.1x, 3.20, 3.30, 3.31, 4.0x, and 4.10, and MASM Versions 1.25, 1.27,
3.0x, 4.00, 5.00, and 5.10.

When you do any floating point arithmetic with the coprocessor all
values are pushed onto the coprocessor stack. The coprocessor stack
only holds 10-byte reals. This means that all real*4 and/or real*8
(real/double) are expanded to 10-byte reals, then all of the
arithmetic is done on these 10-byte reals. The emulator package does
this same expansion. The results of floating point calculations are
then rounded back down to real*4 or real*8 format.

Alternate math calculations are performed using real*4 or real*8
format which can result in less precision than that available with
coprocessor/emulator math.

THE INFORMATION PROVIDED IN THE MICROSOFT KNOWLEDGE BASE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. MICROSOFT DISCLAIMS ALL WARRANTIES, EITHER EXPRESS OR IMPLIED, INCLUDING THE WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL MICROSOFT CORPORATION OR ITS SUPPLIERS BE LIABLE FOR ANY DAMAGES WHATSOEVER INCLUDING DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, LOSS OF BUSINESS PROFITS OR SPECIAL DAMAGES, EVEN IF MICROSOFT CORPORATION OR ITS SUPPLIERS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. SOME STATES DO NOT ALLOW THE EXCLUSION OR LIMITATION OF LIABILITY FOR CONSEQUENTIAL OR INCIDENTAL DAMAGES SO THE FOREGOING LIMITATION MAY NOT APPLY.

Copyright Microsoft Corporation 1986-2002.