Divide Overflow??

Code:

---

mov eax, 1193180 ; a constant
movzx ebx, BYTE PTR [si]
div ebx ;divide overflow here :(

---

I can't figure out how to get around this problem. How can this cause an overflow? Dividing 1193180 by any number from 1 to 255 cause an overflow (I'm filtering out zero)?

This is for a little 16 bit program that is supposed to play a tune on the PC Speaker. The number i'm dividing by is a frequency of the note I want to play (in hertz).

I'm using this site as my guidance on this topic...
http://fly.cc.fer.hr/GDM/articles/sndmus/speaker1.html

How can I get past this problem???

I just read up on using the div instruction since I havent covered that yet (im teaching myself asm). It appears that both edx and eax registers must be initalized first. Here is my example:

Code:

---

.data
        var1 byte 255
.code
        mov edx,0
        mov eax,01193180h
        div var1

---

I should point out that you mentioned this is a 16-bit program, yet you are using 32 bit registers. I assume it matters which ones you use?

For 16 bit replace edx with dx and eax with ax

Quote:

---

Originally Posted by wossname

Code:

---

mov eax, 1193180 ; a constant
movzx ebx, BYTE PTR [si]
div ebx ;divide overflow here :(

---

I can't figure out how to get around this problem. How can this cause an overflow? Dividing 1193180 by any number from 1 to 255 cause an overflow (I'm filtering out zero)?

This is for a little 16 bit program that is supposed to play a tune on the PC Speaker. The number i'm dividing by is a frequency of the note I want to play (in hertz).

I'm using this site as my guidance on this topic...
http://fly.cc.fer.hr/GDM/articles/sndmus/speaker1.html

How can I get past this problem???

---

Well first off you shouldn't be able to use eax in a 16 bit enviroment. Another thing in a 16 bit enviroment, dx is the high part of a dividend. So you have to zero it out before using div.

What compiler are you using?

16 bits i just the addressing system isnt it? I've had no problems using 32 bit registers in realmode for ages. Thats how fast graphics manipulation is perfomed, by shoving data around in 4byte chunks.

I'm on MASM. but this question is way out of date now and I'm no longer interested in it

Quote:

---

Originally Posted by wossname

16 bits i just the addressing system isnt it? I've had no problems using 32 bit registers in realmode for ages. Thats how fast graphics manipulation is perfomed, by shoving data around in 4byte chunks.

I'm on MASM. but this question is way out of date now and I'm no longer interested in it

---

Yea it's memory addressing.