I need my computer architecture hw done

CS-UY 2214 — Homework 2

Jeff Epstein

Introduction

Complete the following exercises. When you finish, submit your solutions on Gradescope. Submit your code files exactly as named in the questions.

Problems

1. If you haven’t yet done so, complete the Verilog tutorial, which you can find on Classes, in the Resources section, in the “guides” folder. You don’t have to submit anything: this is just to familiarize you with the tools.

You may want to look at the Verilog cheat sheet and reference manual, also available on Classes.

2. Download and extract the enclosed files from hw2a.zip:

• full_adder_nodelay.v – a full one-bit adder module • four_bit_adder_subtractor.v – a (broken) four-bit ripple-carry adder/subtractor, which uses

the adder in the above file

• adder_subtractor_test.v – a test bench that will verify the behavior of the above module

To run the test bench, compile adder_subtractor_test.v and run the resulting binary in the simu- lator. View the resulting waveform in GTKWave. The following commands will do it:

iverilog -o adder_subtractor_test.vvp adder_subtractor_test.v

vvp adder_subtractor_test.vvp

gtkwave adder_subtractor_test.vcd

The four_bit_adder_subtractor module should be able to add and subtract four bit numbers. As provided, however, it can currently only add.

The four_bit_adder_subtractor module takes three inputs:

• A – the first input (four bit) • B – the second input (four bit) • Op – the operation (one bit); this will be 0 for addition and 1 for subtraction. Initially, it is

ignored within the module.

And it yields two outputs:

• S – the sum (four bit) • Cout – the carry out (one bit)

1

When you run the Verilog simulator, the test bench will notify you if the adder produces unexpected results. Initially, you should see the following output:

When subtracting 1 and 1, got 2, but expected 0

When subtracting 2 and 5, got 7, but expected -3

When subtracting 4 and -4, got 0, but expected -8

When subtracting 7 and 7, got -2, but expected 0

(The numbers might seem a little funny: the test bench claims that 4 minus −4 should yield −8. This is because in a 4-bit 2’s complement number, it’s impossible to represent 8, the mathemtically correct solution. The bit pattern 1000 will be interpreted as −8.) By modifying the four_bit_adder_subtractor.v file, make it support both addition and subtraction. Do not modify the other files. Your solution must be consistent with the addition algorithm we discussed in class. You may not use Verilog’s built-in arithmetic operators. You may use the conditional operator (?:) as a multiplexer. You can use all the basic gates (&, |, ^, ∼). If you’ve done it correctly, the test bench shouldn’t print out any unexpected results.

Submit your corrected four_bit_adder_subtractor.v file, as well as a screenshot (named four_bit_adder_subtractor.png or four_bit_adder_subtractor.jpg) showing the output of GTKWave when viewing the waveform of ports on the corrected four_bit_adder_subtractor module. Make sure that your waveform doesn’t contain any red (unknown) values.

3. Write a combinatorial Verilog module binary_to_gray. Its input, B, is an unsigned 4-bit binary number. Its 4-bit output, G, is the corresponding gray code. Use only Verilog’s low-level logical operations (operators |, &, ^, ∼) in your answer. Your module with strt like this:

module binary_to_gray(B, G);

Be sure to declare your ports as input or output; declare any additional wires you may need; and use assign statements to connect ports to gates.

One way to convert from binary to gray code is to observe the following rules:

• The MSB (most significant bit) of the output is equal to the MSB of the input. • Bit i of the output is obtained by XOR-ing together bits i and i + 1 of the input (i.e. that bit,

and the one to its left).

Put your code in a file named binary_to_gray.v. Test your code using the binary_to_gray_tb.v test bench file provided in hw2b.zip. Note that the test bench does not automatically verify the test cases; you’ll need to look at the waveform in GTKWave in order to determine if your output is correct. The correct gray codes are given in binary on the Wikipedia page linked above.

Submit your source file (named binary_to_gray.v) and a screenshot of GTKWave showing the wave- forms of your module’s ports (named binary_to_gray.png or binary_to_gray.jpg). Make sure that your waveform doesn’t contain any red (unknown) values.

4. Download the file hw2c.zip.

Using the ALU circuit diagram that we discussed in class as a guide, implement a combinatorial 4- bit Arithmetic Logic Unit in Verilog. Your module must be functionally identical to the unit design presented in class: it must support a 3-bit operation input that selects one of the following operations: A & B, A | B, A + B, A & ∼B, A | ∼B, A - B, A < B. The inputs and output of the unit should be 4-bit values.

2

A circuit diagram of the ALU is available in the zip file, as well as in the class slides.

In your implementation, do not use the high-level Verilog operators +, -, <, etc. Instead, implement the ALU’s logic as given in the diagram, using low-level gate logic (operators |, &, ^, ∼) and an adder. Use the included four_bit_adder module to provide the adder functionality. You should not need more than one instance of this adder. You may use the conditional operator (?:) as a multiplexer. You can nest several instances of the conditional operator to simulate a larger multiplexer.

Use the included alu.v to get started. Test your code using the included Verilog testbench alu_tb.v. Submit your completed alu.v.

Hint: to create an instance of four_bit_adder inside your module, you can use syntax similar to the following. Below, we’re creating an instance of four_bit_adder named the_adder; the name does not matter. We create a 4-bit wire named adder_sum and a 1-bit wire named adder_Cout for the sum and carry out of the adder, respectively, which we connect to the adder. You are responsible for providing reasonable values for A, B, and carry in.

wire [3:0] adder_sum;

wire adder_Cout;

four_bit_adder the_adder(A, B, 0, adder_sum , adder_Cout );

Hint: you can use the ?: syntax to select from one of several options. For example:

// X will be 3 when Y is 0, and 4 otherwise.

assign X = (Y==0) ? 3 : 4;

// Z will be 3 when Y is 0, and 4 when Y is 2, and 5 otherwise.

assign Z = (Y==0) ? 3 : (Y==2) ? 4 : 5;

You already know the OR, AND, and XOR operations as applied to logical (i.e. true/false) inputs. For example, True and False == False.

These operations can also apply to pairs of n-bit binary numbers a and b. In this case, the numbers will be treated as a sequence of bits a(n−1)...0 and b(n−1)...0. The result of a op b will be the sequence of bits arising from applying operation op to pairs of bits from each operand, i.e. an−1 op bn−1, an−2 op bn−2, ..., a0 op b0. When these operations are applied to sequences of bits, they are called bitwise operations.

In C and C++, the & operator is bitwise AND, | is bitwise OR, ^ is bitwise XOR, and ∼ is bitwise NOT. (Distinguish these bitwise operators from their logical/boolean counterparts: && for logical AND; || for logical OR; and ! for logical NOT.)

For example, let’s say we want to calculate 2510 | 510, i.e. the bitwise OR of 25 and 5. That is, in decimal:

2 5 | 5

?

First, let’s convert both numbers to binary. Here, we write 8 bits for each number:

0 0 0 1 1 0 0 1 = 2510 | 0 0 0 0 0 1 0 1 = 510

?

To calculate the answer, we apply the OR operation to the two bits in each column, yielding:

3

0 0 0 1 1 0 0 1 = 2510 | 0 0 0 0 0 1 0 1 = 510

0 0 0 1 1 1 0 1

That value, 000111012, can be converted into decimal:

0 0 0 1 1 0 0 1 = 2510 | 0 0 0 0 0 1 0 1 = 510

0 0 0 1 1 1 0 1 = 2910

Therefore 2510 | 510 = 2910. The unary NOT operation can also be applied bitwise, in which case each of the bits of the input will be

inverted. For example, to calculate ∼ 2510, we convert the value into binary, yielding 000110012. Note that we’ve written the number with 8 bits. Now we invert each bit, giving us 111001102, which is 23010.

For the following questions, use the C/C++ bit-twiddling operators &, |, ^, and ∼. You may also find the operators << and >> helpful.

You don’t need a compiler to verify your work. You should be able to complete the functions based on your own calculations.

The following diagram, showing the terminology used to name individual bits in an 8-bit representation of the decimal number 41, may help:

5. Complete the following C function, which will return a value equal to its parameter, with the third least significant bit flipped on. That is, flipOnBit3(90) should return 94, but flipOnBit3(45) should return 45, since the third least significant bit is already on in 45.

unsigned int flipOnBit3(unsigned int x) {

return YOUR CODE HERE; }

Your solution should be one line. Submit your answer in a plain text file named hw2.txt. Number your answer with the question number.

6. Complete the following C function, which will return a value equal to its parameter, with the second least significant bit toggled, i.e. if the second least significant bit is on, turn it off, and if it is off, turn it on. That is, toggleBit2(90) should return 88, but toggleBit2(45) should return 47.

Hint: for this question, you probably want to use the xor operator (^).

unsigned int toggleBit2(unsigned int x) {

return YOUR CODE HERE; }

Your solution should be one line. Submit your answer in a plain text file named hw2.txt. Number your answer with the question number.

4

7. Complete the following C function, which will return a value equal to its parameter, with the fifth least significant bit flipped off. That is, flipOffBit5(90) should return 74, but flipOffBit5(45) should return 45, since the fifth least significant bit is already off in 45.

Do not make any assumptions about the size of an unsigned int. That is, your solution should work equally well running on an architecture where unsigned ints are 16-bit, 32-bit, or 64-bit.

unsigned int flipOffBit5(unsigned int x) {

return YOUR CODE HERE; }

Your solution should be one line. Submit your answer in a plain text file named hw2.txt. Number your answer with the question number.

8. Your friend Rupert often confuses the C/C++ logical AND operator (&&) with the bitwise AND operator (&). In frustration, he decides to abandon the logical version, and use the bitwise form exclusively. Because C and C++ use zero to represent false, and any nonzero value to represent true, he opines, using the bitwise operator in place of the logical operator will produce equivalent results.

Is Rupert correct to make this substitution? Justify your answer with examples.

5