More Assembly :(

[zapshe]

Sorry for all the assembly! I already hate coding in it.

When I go to divide, it gives me some out of this world answers. This works:

mov ax, 11
mov bl, 37
div bl
mov [bAns16], ah

And will give me the correct remainder of 11. But as soon as I switch those numbers to variables, the whole thing breaks:

mov ax, [bNum1]
mov bl, [bNum4]
div bl
mov [bAns16], ah


;bNum1 = 11
;bNum4 = 37

The answer should be 0 with a remainder of 11. But the answer ends up 104 and with a remainder of 3 somehow.


Moreover, I'm having issues with using imul - makes no sense at all. Same as with div:

mov ax, 33
mov bx, -17
imul ax, bx
mov [wAns15], ax

Will give me the correct answer, but won't as soon as I put variables in there:

mov ax, [bNum5]
mov bx, [bNum6]
imul ax, bx
mov [wAns15], ax

;bNum5 = 33
;bNum6 = -17

So it should equal -561, but gives me -32049 instead consistently.


What am I doing wrong? Assembly hurts me internally.

[deleted account xyzzy]

A personal aside......just commentary, no real advice or help.

Assembly coding is one of the main reasons why I began to learn C++ on my own.

C is bad enough for making simple "oops!" errors, but assembly is like using an elephant gun to swat a fly.

[zapshe]

Figured that out myself. It's pretty easy to understand, but very easy to mess up any little thing

[eugenedakin]

Hello zapshe,

In the first example, I can’t get the correct answer. The reason is 37 is a 16-bit number (bx) and not an 8-bit number (bl), and it should give an error. My error is “operand size conflict” for the first part of your question on Visual Studio 16.4.4 C++ 2019.

INT16 bAns16 = 0;
__asm {
       mov ax, 11
       mov bl, 37
       div bl
       mov [bAns16], ah //error line ‘operand size conflict’
       }

The following code has no errors on my machine and has the correct quotient of 0, and remainder of 11. All numbers are 16-bit.

INT16 bAns16;
INT16 bNum1 = 11;
INT16 bNum4 = 37;
__asm {
       mov ax, [bNum1]
       mov bx, [bNum4]
       div bx
       mov [bAns16], ax
       }

Does this assembly code for the first part of your question work on the computer?

[eugenedakin]

Hi zapshe,

In the second part of the multiplication question, I am guessing that you are attempting to use a 15-bit answer by the use of the variable name wAns15. This needs to be a 16-bit value for the negative value to properly be counted, as the 16th-left-bit is the indicator for a positive or negative sign. The following code on my computer gives me -561 on Visual Studio 16.4.4 C++ 2019.

INT16 bAns16 = 0;
__asm {
       mov ax, 33
       mov bx, -17
       imul ax, bx
       mov [wAns16], ax
       }

The following code has no errors on my machine and has the correct quotient of -561 and requires 16-bit variables.

INT16 wAns16;
INT16 bNum5 = 33;
INT16 bNum6 = -17;
__asm {
       mov ax, [bNum5]
       mov bx, [bNum6]
       imul ax, bx
       mov [wAns16], ax
       }

Do your computer show -561 with this code?

Edit: Changed wording to include indicator, as there is more to it for negative values.

[zapshe]

You declared the variables as INT16, mine are declared as a single byte (as given to us). Assembly sucks. I got a few answers on stackoverflow that helped me solve why imul wasn't working. It was because it was putting the 1byte value in a 2byte container, so it was going into the variable after it to get the data.

I was able to get div to work using the same logic, with either cbw or mov ah, 0:

mov al, byte [bNum1]
;cbw
mov ah, 0
div byte [bNum4]
mov [bAns16], ah

mov ax, [bNum2]
cbw
div byte [bNum3]
mov [bAns17], ah

I'm still working on getting this to work though:

;	bAns18 = wAns13 / bNum2
;	bRem18 = wAns13 % bNum2

mov ax, word [wAns13]
cwd
div byte [bNum2]
mov [wAns18], al
mov [wRem18], ah


;wAns13 is 407
;bNum2 is 15

I've tried everything it seems. I wish assembly could go die in a ditch.

Thanks for the response!

[zapshe]

Thanks for the replies, I was able to solve everything with the information gained here:

https://stackoverflow.com/questions/59997884/assembly-x86-imul-and-div-arent-doing-what-i-expect/59999050?noredirect=1#comment106111904_59999050

[againtry]

That's gratitude for you. In short repliers, you wasted your time.

[eugenedakin]

For the sake of completeness and for those who will look at the cplusplus forums in the future for answers, here is one possible working answer to zapshe's last question:

INT16 wAns13 = 407;
INT16 bNum2 = 15;
INT16 wAns18;
INT16 wRem18;
__asm {
            mov ax, [wAns13]
            div [bNum2]
            mov[wAns18], ax
            mov[wRem18], dx
            }
;wAns18 = 27
;wRem18 = 2

Both wAns18 and wRem18 are INT16 variables, and a 16-bit ax quotient is used with a 16-bit remainder dx.

Edit: Removed cwd as this command is not required.