float binary makeup

[jay vas]

Hi guys , Im trying to figure out how java floats are stored in binary.
The floatToRawIntBits method is supposed to yield an integer that can be masked
to reveal the underlying float representation in terms of (1) exponent and a (2) significand...

However, when i run the following code (to print exponent and significand for numbers 1 through 10) ....

 
for(float val =0; val  < 10 ; val++)
		{
			int i = Float.floatToRawIntBits(val);
			int exp = (i&0x7f800000);//exp
			int sig = (i&0x007fffff);//sig
			System.out.print(Integer.toBinaryString(sig)+" ");
			System.out.println(Integer.toBinaryString(exp));
			
		}

I get some very strange looking binary numbers ... the output is below

0.0 0 0
1.0 0 111111100000000000000000000000
2.0 0 1000000000000000000000000000000
3.0 10000000000000000000000 1000000000000000000000000000000
4.0 0 1000000100000000000000000000000
5.0 1000000000000000000000 1000000100000000000000000000000
6.0 10000000000000000000000 1000000100000000000000000000000
7.0 11000000000000000000000 1000000100000000000000000000000
8.0 0 1000001000000000000000000000000
9.0 100000000000000000000 1000001000000000000000000000000

Can somebody explain this ?

[Jesper de Jong]

The format that Java uses is the single-precision (binary32) format of the IEEE-754 standard.