ECE-C690 Final Exam: Dependable Computing - Error Detection and Correction, Exams of Computer Science

The final exam questions for the ece-c690: dependable computing course, focusing on error detection and correction techniques. The exam covers topics such as error correction using separable 3-of-6 codes, hamming codes, cyclic codes, time redundancy approaches, and checksums in matrix multiplication. Students are required to answer questions related to hardware designs, parity equations, generating polynomials, ripple-carry adders, and data integrity.

Typology: Exams

2012/2013

Uploaded on 05/18/2013

maazi
maazi 🇮🇳

4.4

(12)

75 documents

1 / 4

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
ECE-C690: Dependable Computing
Final Exam
March 13, 2009
The midterm is due in class Monday, March 16, 2009. Please answer all questions. You are
not allowed to collaborate with others.
1. Answer the following questions.
(a) (10 points) A simple technique for error correction is to pick the nearest code word as the correct
word once an error has been detected. Develop a hardware design that performs the error correction for a
separable 3-of-6 code by choosing the nearest valid code word. Make the necessary assumptions and state
them clearly in your answer. Points will be awarded, in part, based on the efficiency of your design.
(b) (15 points) The following is a parity matrix for a Hamming code:
d1c1d2c2d3c3d4
0111010
1001011
1100110
where ciis a check bit and diis a data bit.
Write the parity equations for the three check bits.
Using these parity equations, encode the data word: d1d2d3d4= 0110.
The encoded word 1100001(d1c1d2c2d3c3d4) has a single bit error. Which bit is in error?
(c) (10 points) Consider an n-bit code word and show that if the generating polynomial G(X) of a cyclic
code has more than one term, all single-bit errors will be detected.
1
pf3
pf4

Partial preview of the text

Download ECE-C690 Final Exam: Dependable Computing - Error Detection and Correction and more Exams Computer Science in PDF only on Docsity!

ECE-C690: Dependable Computing

Final Exam

March 13, 2009

The midterm is due in class Monday, March 16, 2009. Please answer all questions. You are not allowed to collaborate with others.

  1. Answer the following questions.

(a) (10 points) A simple technique for error correction is to pick the nearest code word as the correct word once an error has been detected. Develop a hardware design that performs the error correction for a separable 3-of-6 code by choosing the nearest valid code word. Make the necessary assumptions and state them clearly in your answer. Points will be awarded, in part, based on the efficiency of your design.

(b) (15 points) The following is a parity matrix for a Hamming code:

   

d 1 c 1 d 2 c 2 d 3 c 3 d 4 0 1 1 1 0 1 0 1 0 0 1 0 1 1 1 1 0 0 1 1 0

where ci is a check bit and di is a data bit.

  • Write the parity equations for the three check bits.
  • Using these parity equations, encode the data word: d 1 d 2 d 3 d 4 = 0110.
  • The encoded word 1100001(d 1 c 1 d 2 c 2 d 3 c 3 d 4 ) has a single bit error. Which bit is in error?

(c) (10 points) Consider an n-bit code word and show that if the generating polynomial G(X) of a cyclic code has more than one term, all single-bit errors will be detected.

  1. Answer the following questions.

(a) (15 points) Investigate the error detection capability of the following time redundancy approach when used on a ripple-carry adder. During the first addition, the operands are encoded using a 3N arithmetic code. During the second addition, the operands are encoded using a 5N arithmetic code. Will this scheme detect any single error that can occur in the adder, and why? Will this approach detect any double errors that can occur in the adder, and why?

(b) (10 points) Demonstrate the use of checksums in matrix multiplication to detect, locate, and correct an error by multiplying the following two matrices. You must first augment each matrix to include the necessary checksum rows and columns. Show the augmented matrices, the resulting matrix product, and explain how the erroneous value is located and corrected. Assume that element (2, 2) of the matrix product is calculated incorrectly when demonstrating the technique.

A =

B =

  1. (15 points) Consider a computation process that requires 100 hours to execute. The failure rate, i.e., events that create a need to backtrack to the previous checkpoint, is λ per hour. Each checkpoint adds 100 seconds to the computation in way of overhead. Two checkpointing schemes are being considered: (1) inserting 3 checkpoints, one after every 25 hours of computation and (2) inserting a total of 9 checkpoints, one after every 10 hours of computation. Assuming the exponential failure model, determine the values of λ for which one checkpointing scheme would be preferable to the other.