absolute value urgent

[aonks]

can someone please tell me the inline assembly code to find the absolute value of a double variable ??

it is very urgent ...


i'm using GCC in Cygwin

reply soon

[NathanB]

can someone please tell me the inline assembly code to find the absolute value of a double variable ??

it is very urgent ...


i'm using GCC in Cygwin

reply soon

Just bit-wise AND the value with $0FFFFFFF to clear the sign bit.

Nathan.

[aonks]

#include <stdio.h>
#include <stdlib.h>

#define PRECISION 3

double absD (double n)
{

// i need the code here, plz write me the code
// i m due an assignment, and its very urgent

return n;

}

int main (int argc, char **argv)
{
double n = -12.45576;
int number = -1234;

if (argc > 1)
n = atof(argv[1]);

printf("abs(%.*f) = %.*f\n", PRECISION, n, PRECISION, absD(n));


printf("\n\nnumber: %d absolute value: %d\n", number, abs(number));

return 0;
}

[aonks]

plz .. nathan .. i need ur help urgently

do reply .. anything would do

[KingTermite]

It's been years since I've toyed with assembly, but to change sign of a two's compliment number, IIRC, you reverse bits and add 1 (not "just" change sign bit).

so you could (algorithmically)

If ($80000000 AND Number) /* sign bit shows its negative)
Number XOR $FFFFFFFF
Number = Number +1

[KingTermite]

Just bit-wise AND the value with $0FFFFFFF to clear the sign bit.

Nathan.

Numbers should be in two's compliment, so its more than "just" changing the sign bit, IIRC.

Also, you show that in Hex, so you would be clearing the top 4 bits. For your example, wouldn't you want to AND it with $7FFFFFFF?

[rye_]

Asking for advice is one thing, or introducing a discussion, but people aren't going to do your homework for you. It just not on.

Ryan

[aonks]

i 've no clue about the syntax either .. i know, i'm bein a pain in da *** .. but plz could u tell me the exact code .. my prof. is gonna flunk me other wise

[mike_g]

Just bit-wise AND the value with $0FFFFFFF to clear the sign bit.

Nathan.
Wouldent a double consist of 8 bytes? Also, to mask out the sign bit on a 32 bit integer would be 0x7FFFFFFF. Never really did any bitwise stuff with floating point tho.



// i need the code here, plz write me the code
// i m due an assignment, and its very urgent

What have you tried?

[Phenax]

Haven't tried it:


cltd
xorl %edx,%eax
subl %edx,%eax

[aonks]

m really sorry for this .. i know, i should have studied, this is not like me .. but can u help me, just this one time

[KingTermite]

Never really did any bitwise stuff with floating point tho.
Just realized when you said this that he was asking for a double value. That's a WHOLE DIFFERENT ball game. I was thinking of a 2's compliment integer.

I think you can just change the sign bit (which is the 1st bit).

Double =
1-11-52 (bit fields for sign-exponent-mantissa) IIRC.

Just check the first bit (double is 64-bit) and change to a 0. Actually...you can just force to a 0 and not worry about checking.

Number = Number AND $7FFFFFFFFFFFFFFF

[aonks]

phenax .. can u tell me the exact syntax ?

[aonks]

i can understand your logic .. but have no clue about the syntax

[KingTermite]

I think you've had enough help....if you know your assembly at all, you should be able to this command or two easily. If you can't even do that command or two, then quite frankly, you should flunk.

[aonks]

plz ... i beg of u
dnt ditch me ... i 'll
have 2 repeat my whole semester

[Phenax]

Actually:
http://www.scriptlance.com/projects/1212744106.shtml

He's signed up to create it for a customer at ScriptLance for $30. Do your own work.

[KingTermite]

Actually:
http://www.scriptlance.com/projects/1212744106.shtml

He's signed up to create it for a customer at ScriptLance for $30. Do your own work.

LOL....nice find!

[mike_g]

Rofl!

[aonks]

mr. smarty pants ...

[NathanB]

Numbers should be in two's compliment, so its more than "just" changing the sign bit, IIRC.

Also, you show that in Hex, so you would be clearing the top 4 bits. For your example, wouldn't you want to AND it with $7FFFFFFF?

Yeah, that's it. Nice catch!

Nathan.

[KingTermite]

Yeah, that's it. Nice catch!

Nathan.

That's why we're here as a community. :) :popcorn:

[aonks]

fine .. u guyz busted me, i'm guilty as charged. but plz, help out this poor lad .. for some charity work .. god will bless u all

[slavik]

May god be with you on your journey!

[aonks]

i might sound stupid, but am not dat dumb .. dis looks lyk the code for an infinte loop

[Lster]

It is. Do not run the command above!

[LaRoza]

i might sound stupid, but am not dat dumb .. dis looks lyk the code for an infinte loop

This forum is for Ubuntu support. It has other sections that are not directly related to Ubuntu, but none of them are to be used this way.

I see all your posts are in this thread. This forum isn't here to do your work, but to help you.

Please do not abuse this forum. If you have no interest in the topics of this forum, please don't use it. (I am going to help with this ;))