Conversion of float to byte value in C language

[Parag Ambernath Keni]

Hi,
I am using the Diab compiler for the MPC 555 controller. I wanted to know the internal bit representation of float value. Also if I type cast of float value to byte the way the final value that will be computed.

The code snippet is as follows:

typedef float F32;
typedef unsigned char U8;
typedef unsigned long int U32;

U32 my function(void)
{
U32 cs_sum = 0;
F32 cs_data = 50.157;

cs_sum = (U8) (cs_data);
return cs_data;
}

[sicarie]

Hi,
I am using the Diab compiler for the MPC 555 controller. I wanted to know the internal bit representation of float value. Also if I type cast of float value to byte the way the final value that will be computed.

The code snippet is as follows:

typedef float F32;
typedef unsigned char U8;
typedef unsigned long int U32;

U32 my function(void)
{
U32 cs_sum = 0;
F32 cs_data = 50.157;

cs_sum = (U8) (cs_data);
return cs_data;
}

I'm not sure I understand your code or the second part of your question, but I think - please someone correct me if I'm wrong (which I have been more often than not lately - I'm rusty on my C) - that you can create the float you would like to see, and then use C's printf() command. You can use the first part of the argument to set the value to whichever representation you want, and then give it your float variable after (though you might have to cast before you use printf(), I can't remember).

[RRick]

The internal representation of the float or double is system dependent. You might find some info in the system or compiler include files. For linux, /usr/include/limits.h and some math includes under /usr/include/bits have info.

Since the format of floats and ints are radically different, usually you let the compiler do the conversions. Your example would convert a float to a U8 and then put the results into a U32. The end result would be 0...255, and this sounds like you're going to loose info.

Usually float to int conversions drop the fractional part and keep the whole integer. Some systems use rounding. You also have the problem of putting something too large into U8 or U32. Once again it's system dependent what happens.

Try a few tests and see what happens.