Alternate Math (BC /FPa) Won't Always Produce Smaller .EXE's (58035)

SUMMARY

Using alternate math (BC /FPa) does not always produce smaller executable files than using the emulator math library (BC /FPi) when compiling with Microsoft Basic Compiler Versions 6.00 and 6.00b for MS-DOS and MS OS/2 and Microsoft Basic Professional Development System (PDS) Versions 7.00 and 7.10 for MS-DOS and MS OS/2.

Page 8 of the "Microsoft Basic Compiler 6.0: User's Guide" and Page 562 of the "Microsoft Basic 7.0: Programmer's Guide" (for 7.00 and 7.10) misleadingly state that "it [the alternate math library] also creates a smaller executable file." Page 545 of the "Microsoft Basic 7.0: Programmer's Guide" misleadingly states that "alternate math is an IEEE-compatible math package optimized for speed and size."

Although there is an initial space savings from using the alternate math libraries, each individual floating-point calculation can take more room using the alternate math library than the equivalent code for the emulator math library. This means that as the code grows in size, the initial space savings can be lost and the program can actually be larger using the alternate math package.

NOTE: You will only notice a savings in the size of an executable compiled /FPa versus /FPi if the program is also compiled with the /O (stand-alone) option. If you compile as non stand alone with the alternate math library (/FPa) option, the program will actually contain both math libraries -- the compiled program will contain the alternate math routines, while the Basic run-time module (BRUN60Ax.EXE, BRUN61Ax.EXE, or BRT70Axx.EXE) will contain the emulator math routines.

MORE INFORMATION

The small program below demonstrates the size difference when compiling with and without the alternate math package.

Compile the following program in these two ways:

BC /FPa test, test1.obj;
BC /FPi test, test2.obj;
LINK test1;
LINK test2;

Code Example

   REM  This is TEST.BAS
   a = 6.1
   b = 5.7
   c = a * b

The program compiled with the alternate math package (/FPa) produces larger code for the floating-point operation than the equivalent instructions using the emulator math library (/FPi). There is, however, an initial savings of about 4K when compiling with the alternate math library. In most programs, this initial savings offsets any extra code generated by floating-point calculation. In larger programs with lots of floating-point calculations, the extra code for each floating-point operation can actually create a larger executable file.