Programming Forums - Floats to doubles

Programming Forums (http://www.programmingforums.org/forumindex.php)

- Assembly (http://www.programmingforums.org/forum20.html)

- - Floats to doubles (http://www.programmingforums.org/showthread.php?t=6874)

Klarre

Nov 7th, 2005 1:30 PM

Floats to doubles

I have a problem with floats and doubles.

I wanted to put a camera in my OpenGL scene. But when
I added the gluLookAt function, the program did'nt started
anymore.

I was using this code, and simply made a call like

_gluLookAt eyex, eyey, eyez, centerx, centery, centerz, upx, upy, upz

in my draw function.
These variables are created like this:

eyex dd 0.0

_gluLookAt macro A1, A2, A3, A4, A5, A6, A7, A8, A9

        push A9

        push A8

        push A7

        push A6

        push A5

        push A4

        push A3

        push A2

        push A1

        call gluLookAt

endm

I guessed that the problem was that the gluLookAt function expects 64 bits doubles,
instead of 32 bits floats. I tried my assumption by pushing twice as much on the stack.
Then the program started again, but with wrong values to gluLookAt of course.

How shall I solve this problem? How do I convert my floats to doubles? Is there
a simple way of doing this?

I am not an assembly guru, so try keeping the answears as clean as possible ;)

Thanks

/Klarre

Polyphemus_

Nov 7th, 2005 1:36 PM

Maybe this page helps: http://www.osdata.com/topic/language/asm/convert.htm

Good luck :)

Klarre

Nov 7th, 2005 2:39 PM

Thanks for the link! It took me back on track again.

But...

I got that I should use the cdq instruction to extend the eax to edx. Right?

Am I thinking correct if I try using this rewritten code instead of the old one?

_gluLookAt macro A1, A2, A3, A4, A5, A6, A7, A8, A9

        mov eax, A9

        cdq

        push eax

        push edx



        mov eax, A8

        cdq

        push eax

        push edx



        ...



        mov eax, A1

        cdq

        push eax

        push edx



        call gluLookAt

endm

Klarre

Nov 7th, 2005 4:12 PM

Problem solved...

iignotus

Nov 7th, 2005 6:04 PM

Mind posting the solution to help others?

Klarre

Nov 8th, 2005 5:09 AM

I solved it using constants. It is a solution so far. And next step is to use double variables instead, so I can move the camera around in real time.

The original function used

void gluLookAt(GLdouble eyeX, GLdouble eyeY, GLdouble eyeZ,

               GLdouble centerX, GLdouble centerY, GLdouble centerZ,

               GLdouble upX, GLdouble upY, GLdouble upZ);

In C++ the code would look like this:

double eyez = 45.0;

double upy  = 1.0;



gluLookAt(0, 0, eyez, 0, 0, 0, 0, upy, 0);

Rewritten in assembly it looks like this:

_45d0 equ 40468000h  ; eyez

_45d1 equ 0

_1d0  equ 1072693248 ; upy

_1d1  equ 0



;; upx, upy, upz pushed on the stack

push 0

push 0

push _1d0

push _1d1

push 0

push 0



;; centerx, centery, centerz pushed on the stack

push 0

push 0

push 0

push 0

push 0

push 0



;; eyex, eyey, eyez pushed on the stack

push _45d0

push _45d1

push 0

push 0

push 0

push 0



call gluLookAt

All times are GMT -5. The time now is 4:24 PM.