Converting uint to float in C++

Question

I'm using ReadProcessMemory in order to get 4 bytes. The function allows me to represent this as an unsigned int. I wish to represent this as a float; or in other words use the byte representation of this uint for my float. I've tried casting and it does not seem to work.

Example: byte representation: 94 4E 2D 43

uint: 1127042708

float: 173.3069458..

Any help would be appreciated.

Expand on "does not seem to work" as that is the correct solution. — Yann Ramin, Feb 20 '11 at 13:58
How do you cast the uint value to float. Do you use a static or a reinterpret cast? — Fox32, Feb 20 '11 at 14:01
See http://stackoverflow.com/questions/3991478/building-a-32bit-float-out-of-its-4-composite-bytes-c — Sergei Tachenov, Feb 20 '11 at 14:08

score 5 · Accepted Answer · edited Feb 20 '11 at 14:13

5

ReadProcessMemory() takes a pointer to void so it is up to you to point it at a float.

float f;
ReadProcessMemory(hProcess, lpBaseAdress, &f, sizeof(f), NULL);

Note that this will break when sizeof(unsigned int) != sizeof(float).

Casting won't work because it will take the value and not the representation of the integer.

edited Feb 20 '11 at 14:13

Matteo Italia

123,740
17
206
299

answered Feb 20 '11 at 14:06

markus

495
4
5

+1, but why should it break when `sizeof(unsigned int)!=sizeof(float)`? We are telling to `ReadProcessMemory` to read exactly the size of a `float`, if the data read from the other process starts as a `float` (otherwise the whole thing wouldn't make sense :) ) it shouldn't break. – Matteo Italia Feb 20 '11 at 14:16
I don't think there is a sizeof problem on platforms where ReadProcessMemory actually exists. This code is not very portable, is it? :-) – Bo Persson Feb 20 '11 at 14:33

score 4 · Answer 2 · edited Feb 20 '11 at 14:04

4

UINT d= 0x432D4E94;
float f= *(float*)&d; // 173.30695

edited Feb 20 '11 at 14:04

Sergei Tachenov

24,345
8
57
73

answered Feb 20 '11 at 14:01

Lior Kogan

19,919
6
53
85

score 4 · Answer 3 · answered Feb 20 '11 at 14:05

The safest way (that takes in account possible alignment problems) is to use a union

#include <stdio.h>

int main(int argc, const char *argv[])
{
    union FloatOrUInt
    {
        float asFloat;
        unsigned int asUInt;
    } fu;

    fu.asUInt = 1127042708;
    printf("Float value = %0.6f\n", fu.asFloat);

    return 0;
}

Note however that even if you know that floats are in standard IEEE754 format there can be problems for endianness.

score 2 · Answer 4 · answered Feb 20 '11 at 14:03

2

unsigned int input = 1127042708;

float output = *reinterpret_cast<float*>(&input)

or

float output = *(float*)(&input)

answered Feb 20 '11 at 14:03

Fox32

13,126
9
50
71

Converting uint to float in C++

4 Answers4

Linked