At 09:06 AM 05/08/2000 -0700, you wrote:
>>If I do the following
>>$CEE1289A $100 / h.
>>>$FFCEE129
>>
>>but if I perform the following:
>>$CEE1289A 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ h.
>>>$FFCEE128

>I don't know what the correct answer is. Can anyone run this on
>another 32-bit Forth and report the results?
>
>Thanks,
>
>Rick

Win32For:
        $CEE1289A $100 / h. FFCEE128 ok

Gforth:
        $CEE1289A $100 / hex u. FFCEE129 ok

It's getting interesting ;)

Win32For is in assembler; / uses /mod, and /mod is:

CODE /MOD ( n1 n2 -- rem quot )
                pop ecx
                mov eax, ecx
                xor eax, ebx
                jns short @@1

                mov eax, ecx \ low order part to eax
                mov ecx, edx \ save UP
                cdq
                idiv ebx
                test edx, edx \ set zero flag
                je short @@2

                add edx, ebx \ add divisor to remainder
                dec eax \ decrement quotient
                jmp short @@2

@@1: mov eax, ecx \ low order part to eax
                mov ecx, edx
                cdq
                idiv ebx
@@2: push edx
                mov ebx, eax
                mov edx, ecx \ restore UP
                next c;

Gforth is in C, and / uses, as one would expect, the C division:

NAME("/")
{
DEF_CA
Cell n1;
Cell n2;
Cell n;
NEXT_P0;
n1 = (Cell) sp[1];
n2 = (Cell) TOS;
sp += 1;
{
n = n1/n2;
}
NEXT_P1;
TOS = (Cell)n;
NEXT_P2;
}

        If you regard the numbers as signed, the Gforth/SwiftForth result is
correct. If they are unsigned, Win32For is correct.

        $CEE1289A is -824104806

        $FFCEE128 is -3219160
        $FFCEE129 is -3219159

On a calculator:
        -824104806 / 256 = -3219159.3984375

I don't know what the standard result should be.

        Elko

-----------------------------------------------------------------------
I want to be a nothing-knower, Elko Tchernev
a little ant on any hill; etcher1_at_umbc.edu
for time is dead, the sun is over www.gl.umbc.edu/~etcher1
and there is nothing left to kill. tel/fax (410) 964 0593

.
Received on Mon May 08 2000 - 13:09:51 PDT