how to write negative float value in MASM?

[supercoollee]

i know i can predefine some float values in .data
but if i want to input instant value, how should i write it?

i tried this:  invoke functionname,3f800000h,0,bf800000h

here 3f800000h is interpreted into hex 3f800000, but bf800000h can't be assembled.

i also tried -3f800000h, a wrong value was produced. how should i write the right value?

i also tried invoke functionname,1.0,0,-1.0   . this can't be assembled either.

[dedndave]

when hexidecimal values begin with a non-numeric character, they must be preceeded with a 0: 0bf800000h
otherwise, the assembler thinks it is  a label of some sort
labels, on the other hand, may not begin with a numeric digit

[GregL]

INVOKE can't accept floats directly like that. MASM32 includes some macros for doing what you want called FP4, FP8 and FP10. You use them like this: 

  INVOKE functionname, FP8(1.0), FP8(0.0), FP8(-1.0)

[joemc]

i didn't know you could write negative numbers in hex using the minus sign :) i thought you just had to make sure the negative bit was set. ... but wikipedia says you can and they are very reliable :)  why do you want to represent a negative number in hex?

[GregL]

From the MASM Programmer's Guide:

; Encoded as hexadecimals
ieeeshort       REAL4    3F800000r             ; 1.0 as IEEE short
ieeedouble      REAL8    3FF0000000000000r     ; 1.0 as IEEE long
temporary       REAL10   3FFF8000000000000000r ; 1.0 as 10-byte real

I've never done it that way, but I guess it's possible.  It doesn't show an example of negatives.

[dedndave]

it is very time-consuming to do it that way, but it can be done
for floats, there is no 2's compliment - just take the same positive value and set the sign bit
why do that way when the macros are there for ya - lol
if you do it by hand, you are more likely to make a mistake

[supercoollee]

i didn't know i could write:
INVOKE functionname, FP8(1.0), FP8(0.0), FP8(-1.0) . now this is great!

i wrote hex only because invoke functionname,1.0,2.0,3.0  can't be assembled.
and some float values are very easy to remember , such as 0, 0.5, 1 and -0.5, -1

[Farabi]

MASM Included a macro for FPU, CFLT and CDBL.

[supercoollee]

i find INVOKE functionname, FP8(1.0), FP8(0.0), FP8(-1.0)  can't help the situation when a function PROTO is like this:
function PROTO :dword,:dword,:dword,:dword,:dword,:dword

i still have to predefine 3 double values in .DATA , then
invoke function dword ptr predefined1,dword ptr predefined1+4,dword ptr predefined2,dword ptr predefined2+4,dword ptr predefined3,dword ptr predefined3+4

and FP4(1.0) is actually assembled the same way as if a float value was predefined in .DATA .

[MichaelW]

Try:

function PROTO :REAL8,:REAL8,:REAL8

or

function PROTO :QWORD,:QWORD,:QWORD

[jj2007]

and FP4(1.0) is actually assembled the same way as if a float value was predefined in .DATA .

Not by accident:

Code Select Expand

      FP4 MACRO value
        LOCAL vname
        .data
        align 4
          vname REAL4 value
        .code
        EXITM <vname>
      ENDM


\masm32\macros\macros.asm is a fascinating lecture :wink

[supercoollee]

Try:
function PROTO :REAL8,:REAL8,:REAL8
or
function PROTO :QWORD,:QWORD,:QWORD

i didn't mean the protos which can be defined by me, i meant the predefined protos such as

gluPerspective proto :dword,:dword,:dword,:dword,:dword,:dword

this function requires 3 real8 value, but in opengl32.inc , it requires 6 dwords  :tdown

[supercoollee]

and FP4(1.0) is actually assembled the same way as if a float value was predefined in .DATA .

Not by accident:

Code Select Expand

      FP4 MACRO value
        LOCAL vname
        .data
        align 4
          vname REAL4 value
        .code
        EXITM <vname>
      ENDM


\masm32\macros\macros.asm is a fascinating lecture :wink

now that i find out this fact, i will stick to my way - writing immediate HEX value in the code, that's faster than FP4(1.0) in runtime, although the difference is small.

[oex]

It will likely be easier for you to use FP4/8/10/etc(x.x)

Reason being that once you start entering 'normal' numbers you dont have to convert.... ie point:

What is 608.735629 in HEX in the time it takes you to type FP4(608.735629)?

Further when you read that number in hex how long does it take you to know what the value is

[oex]

i didn't mean the protos which can be defined by me, i meant the predefined protos such as

gluPerspective proto :dword,:dword,:dword,:dword,:dword,:dword

this function requires 3 real8 value, but in opengl32.inc , it requires 6 dwords  :tdown

You'll have to ask Farabi on this.... I think he uses glu.inc instead.... I wrote navigation some more complex way without which I still dont completely get :lol