sample paper for dsd

2
Center for Advanced Studies in Engineering Mid Exam CE-3201 DIGITAL SYSTEM DESIGN Time: 90 mins Instructor: Syed Ali Jabir Marks: 50 Date: 20 th July, 2012 Q1. In hardware, subtraction is also done using an adder by adding 2s complement of the number to be subtracted i.e. − = + ( + ) 1. Design a custom adder that performs all the given operations (using only one adder) A + B A – B -A + B -A – B 2. Write RTL Code for your design. 3. TAKE HOME: Write the code on a separate sheet and submit the softcopy of the running code with stimulus. You can ONLY make minor changes to correct the syntax. Marks (8, 7, 5) Q2. float x[5] = {0.22,-0.135,0.877,-0.516,-0.99}; float h[3] = {-2.375,1.24,0.531}; Convolution is to be done on the numbers declared in C/C++ code. Define an appropriate 16-bit fixed point format for the given arrays and convert them into fixed point. What will be the Q format of y[n]? [] = ∑ [][ − ] = Marks (10)

Upload: mtkhusro92

Post on 22-Nov-2015

3 views

Category:

Documents


0 download

DESCRIPTION

DIGITAL DESIGN QUESTIONS

TRANSCRIPT

  • Center for Advanced Studies in Engineering

    Mid Exam

    CE-3201 DIGITAL SYSTEM DESIGN Time: 90 mins Instructor: Syed Ali Jabir

    Marks: 50 Date: 20th July, 2012

    Q1.

    In hardware, subtraction is also done using an adder by adding 2s complement of the number

    to be subtracted i.e.

    = + ( + )

    1. Design a custom adder that performs all the given operations (using only one adder)

    A + B

    A B

    -A + B

    -A B

    2. Write RTL Code for your design.

    3. TAKE HOME: Write the code on a separate sheet and submit the softcopy of the running code

    with stimulus. You can ONLY make minor changes to correct the syntax.

    Marks (8, 7, 5)

    Q2.

    float x[5] = {0.22,-0.135,0.877,-0.516,-0.99};

    float h[3] = {-2.375,1.24,0.531};

    Convolution is to be done on the numbers declared in C/C++ code. Define an appropriate 16-bit

    fixed point format for the given arrays and convert them into fixed point. What will be the Q

    format of y[n]?

    [] = [][ ]

    =

    Marks (10)

  • Q3.

    Consider the above hypothetical DFG

    1. Compute buffer sizes to implement the DFG by a greedy approach.

    2. Compute Buffer sizes to implement the DFG so that no single node has to wait.

    3. Map the DFG on hardware using sequential approach.

    Marks (3, 4, 3)

    Q4.

    Add the following Q format numbers with using sign extension logic (compute CV).

    1111010 Q3.4

    100101 Q5.1

    01110100 Q1.7

    Marks (10)