Remember
  B = 8b      (one byte is 8 bits)
  G = 2^30
  M = 2^20
  k = 2^10

Problem 1
  16G * 32  == (16*32)G = 512G
  8M / 32 == (8*1024*k) / 32 = 8192k/32 = 256k
  log2(2G) = log2(2^31) = 31

Problem 2
  8GB/16b = 8GB/2B = 4G
  4M*32b/B = 4M*(32b/8b) = 4M*4 = 16M
  log2(4M) = log2(2^22) = 22

Problem 3
  26 = 0b11010 --> 00011010.00000000 --> 0001101000000000
  0.3 = 0b0.010011001100110011(0011)* 
      =~ 00000000.01001101 [rounded] --> 000000001001101
         000000001001100 --> 0.29687500   [0.0104 error]
         000000001001101 --> 0.30078125   [0.0026 error]
  0.111111111111111..... = 0b0.000111(000111)* 
      =~ 00000000.00011100 [rounded] --> 0000000000011100

Problem 4
  1111000011110000 = 11110000.11110000 =  -16 + 15/16  =  -15.0625
  0000111100001111 = 00001111.00001111 =   15 + 15/256 =   15.05859375
  0000100000000000 = 00001000.00000000 =    8

Problem 5
  26.5 --> 11010.1 --> 1.10101 * 2^4
    sign --> 0
    exp = 127 + 4 = 131 --> 10000011
    1.10101 -->

 0 10000011 10101000000000000000000 -->  01000001110101000000000000000000
                                    -->  0x41D40000

Problem 6
  0x41540000 -->
     0100 0001 0101 0100 0000 0000 0000 0000 -->
     0 10000010 10101000000000000000000 -->
     +    130         1.101010
     +      3         1. + 0.5 + 0.125 + 0.03125
     1   *  8    *    1.65625
               13.25
  Yep.... by the change of one bit, it's half of 26.5