Department of Electrical and Computer Engineering
The University of Texas at Austin
EE 306, Fall 2006
Problem Set 1
Due: 13 September, before class
Yale N. Patt, Instructor
TAs: Aseem Bathla, Cameron Davison, Lisa de la Fuente, Phillip Duran, Jose Joao,
Jasveen Kaur, Rustam Miftakhutdinov, Veynu Narasiman, Nady Obeid, Poorna Samanta
Instructions:
You are encouraged to work on the problem set in groups and turn
in one problem set for the entire group. Remember to put all
your names on the solution sheet. Also, remember to put the name
of the TA and the time for the discussion section you would like the problem
set turned back to you. Show your work.
Please remember to turn in the Student Information Sheet with a recognizable photograph on the same day you turn in this problem set.

(Adapted from problem 1.5 in the textbook)
Say we had a "black box," which takes two numbers as input and outputs their sum. See Figure 1.7a in the Textbook. Say we had another box capable of multiplying two numbers together. See figure 1.7b. We can connect these boxes together to calculate p * (m + n). See Figure 1.7c. Assume we have an unlimited number of these boxes. Show how to connect them together to calculate:

ax+b

The average of the four input numbers w, x, y, and z

a^{2} + 2ab + b^{2} (can you do it with one add box and one multiply box?)

a^{6} (can you do it using only 3 multiply boxes?)

(1.14)
Suppose we wish to put a set of names in alphabetical order. We call the act of doing so sorting. One algorithm that can accomplish that is called the bubble sort. We could then program our bubble sort algorithm in C, and compile the C program to execute on an x86 ISA. The x86 ISA can be implemented with an Intel Pentium IV microarchitecture. Let us call the sequence "Bubble Sort, C program, x86 ISA, Pentium IV microarchitecture" one transformation process.
Assume we have available four sorting algorithms and can program in C, C++, Pascal, Fortran, and COBOL. We have available compilers that can translate from each of theses to either x86 or SPARC, and we have available three different microarchitectures for x86 and three different microarchitectures for SPARC.

How many transformation processes are possible?

Write three examples of transformation processes.

How many transformation processes are possible if instead of three different microarchitectures for x86 and three different microarchitectures for SPARC, there were two for x86 and four for SPARC.
 (2.3)

Assume that there are about 400 students in your class. If every student is
to be assigned a unique bit pattern, what is the minimum number of bits required to do this?

How many more students can be admitted to the class without requiring additional bits for each student's unique bit pattern?
 (Adapted from 2.13)
Without changing their values, convert the following 2's complement binary numbers into 8bit 2's complement numbers.
 011010
 1001
 1111111000
1 001
 (2.22)
Create two 16bit 2's complement integers such that their sum causes an overflow.
 (Adapted from 2.37) This problem has been postponed. A modified version will appear on a future Problem Set.
If n and m are both 4bit 2's complement numbers, and s is the 4bit result of adding them together, how can we determine, using only the logical operations described in Section 2.6, if an overflow has occurred during the addition? Develop a "procedure" for doing so. The inputs to the procedure are n, m, and s, and the output will be 0 if no overflow occurred and 1 if an overflow did occur.
 (2.42)
A computer programmer wrote a program that adds two numbers. The programmer ran the program and observed that when 5 is added to 8, the result is the character m. Explain why this program is behaving erroneously.
 (Adapted from 2.50)
Perform the following logical operations. Express your answers in hexadecimal notation.
 xFEDC OR x9876
 x1234 XOR x1234
 xFEED AND (NOT(xBEEF))
 (2.54) This problem has been postponed. It will appear on a future Problem Set.
Fill in the truth table for the equations given. The first line is done as an example.
Q1 = NOT (NOT(X) OR (X AND Y AND Z))
Q2 = NOT ((Y OR Z) AND (X AND Y AND Z))

(2.55 ad)
We have represented numbers in base2 (binary) and in base16 (hex). We are now ready for unsigned base4, which we will call quad numbers. A quad digit can be 0, 1, 2, or 3.

What is the maximum unsigned decimal value that one can represent with 3 quad digits?

What is the maximum unsigned decimal value that one can represent with n quad digits (Hint: your answer should be a function of n)?

Add the two unsigned quad numbers: 123 and 121.

What is the quad representation of the decimal number 42?