Discussion #2 Notes
01/28/08
Practice Midterm Problem #1 Write the CUSP floating point representation for the decimal 135.0625.
Remember the CUSP’s representation is 8-bit exponent, followed by the sign bit, followed by the 16-bit
mantissa which excludes the leading 1 (leading one is implicit). Thus, as 13510 = 100001112.062510 =
.00012. The binary representation of 135.0625 is 10000111.0001 = 1.000011100010000 ∗27
8-bit exponent field = 7 = 135excess128 = 100001112Sign bit = 0, positive number 16 bit mantissa field with
implicit leading 1 = 000011100010000
Putting all that together you get: 100001110000011100010000 which is 870710.
Practice Midterm Problem #2 Signed (2scomplement) values are added or substracted in exactly the
same manner as unsigned ones. Remember, as mentioned in the book, that computers do not substract –
they negate the second operand and add them together. Also, in the case of unsigned addition, overflow
occurs when the correct sum is too large to be represented using the available bits. In the case of unsigned
substraction, overflow occurs if the value being substracted is the larger one. In the case of signed addition
as well as substraction, overflow occurs when the result has an incorrect sign.
Assume signed, 2scomplement operation.
A56FF3
+ 9FF7FF
OV 456782
381734
- 808080
B796B4
A689BE
- B8FDE3
A689BE
+ 47021D
ED8BDB
Exercise #1
1. What is the purpose of mini os.lst and mini os.obj files?
2. Say that you are writing a program and that it currenty contains a bug. How would you go about
debugging it?
1
