i'm trying to create a small nasm program which do this operation in floating point
while(input <= 10^5) do
begin
input = input * 10
i = i - 1
end
the equivilant program in nasm is as following
section .data
input: resd 1
n10: dd 0x41200000 ; 10
_start:
mov eax, 0x43480000 ; eax = 200
mov dword [input], eax ; input = eax = 200
mov edx, 0x49742400 ; 10^5
; %begin
mov ecx, 0 ; i = 0
jmp alpha
alpha:
cmp [input], edx ; input <= 10^5
jle _while
jmp log2
_while:
fld dword [input] ; input
fmul dword [n10] ; input * 10
fst dword [input] ; input = input
dec ecx ; i = i - 1
jmp alpha
the _while loop is iterating infinitely
ecx / i gards always the same value = 0 (it is sepposed to be 0) and doesn't decrement
This works for me (tested in DosBox):
org 0x100
bits 16
_start:
mov dword [input], __float32__(99.0)
mov edx, __float32__(10000.0)
mov ecx, 0 ; i = 0
jmp alpha
alpha:
cmp [input],edx ; input <= 10^5
jle _while
jmp log2
_while:
fld dword [input] ; input
fmul dword [n10] ; input * 10
fstp dword [input] ; input = input
inc ecx ; i = i - 1
jmp alpha
log2:
; print the value of cl
mov dl,cl
add dl,'0'
mov ah,2
int 21h
; Exit to DOS
mov ah,0x4c
int 21h
n10: dd 10.0
input: resd 1
Note the bits 16 which tells nasm that 16-bit operands are the default and that instructions using 32-bit operands should be prefixed. Without this your code would be seen as gibberish if you try to execute it in a real-mode environment.
It's possible that you'll need to use bits 32 instead, depending on your target environment.
Also note the use of floating-point literals rather than hex values (you had a typo in your code where you compared against 10^6 instead of 10^5).