浮点到字节数组转换

Question

We use

float f= 3.5f ;
BitConverter.GetBytes(f);

It returns byte array of length 4. But I wonder the math behind this. Anyone will teach the math behind this conversion? Also double conversion info is appreciated.

Answer 1

Under the covers, it is using unsafe, C-style pointers to copy the underlying 32-bit value into a 32-bit array (a byte[4] ):

int rawBits = *(int*)&value;
byte[] bytes = new byte[4];
fixed(byte* b = bytes)
    *((int*)b) = rawBits;
return bytes;

The results are architecture dependent, insofar as the order of the bytes matches the memory byte order (endianness) of the machine where the copy took place.

For example, if the 32-bit rawBits value is 0xAABBCCDD , and it is represented in memory as AA BB CC DD , then the array will contain 0xAA, 0xBB, 0xCC, 0xDD . If the in-memory representation was DD CC BB AA , then the array will contain 0xDD, 0xCC, 0xBB, 0xAA .