Testing Math Coprocessor Instructions FLD FADD FSTP FST (59127)

SUMMARY

The following code was used to view coprocessor instructions on an 80386 machine with an 80387 coprocessor. It is a combination of the examples on Pages 384 and 385 of "Microsoft Macro Assembler 5.1 for the MS-DOS Operating System Programmer's Guide."

               .model small
               .386
               .387
               .stack 100h
               .data
      m1       dd   1.0
      m2       dd   2.0
               .code
      start    proc
               mov ax, @data ;load address of segment
               mov ds,  ax   ;     register
               fld   m1      ;push m1 into st(0)
               fld   m2      ;push m2 into st(0)
               fadd  m1      ;add m2 to st(0)
               fstp  m1      ;pop st(0) into m1
               fst   m2      ;copy st(0) into m2

               fld1          ;push 1 into st(0)
               fldpi         ;push pi into st(0)
               fadd          ;add st(0) and st(1)
                             ;and pop, place result in st(0)
      start    endp
               end  start
				

While assembling the above code, the .386 and .387 directives must be placed below the .model directive to generate 16-bit segments. If placed above the .model directive, 32-bit segments will be generated and will produce erroneous results, as 32-bit segments cannot be run under DOS.

While running the code through CodeView, the number 7 entered in the dialog box displays a dump of the present status of the chip.