UNIT 1
Q1. How floating point numbers are represented in computer systems?
Q2. What do you mean by normalized floating point representation of numbers?
Q3. Why and where the normalization of floating points numbers is recommended?
Q4. Write a short note on floating point representation and arithmetic.
Q5. Describe the consequence of performing various arithmetic operations with normalized
floating point numbers, by taking examples.
Q6. Briefly enumerate pitfalls in floating point arithmetic operations. Use appropriate examples
in support of your answer.
Q7. Explain with the help of a suitable example the operation multiplication and addition on two
numbers represented in normalized floating point notation.
Q8. What are the consequences of normalization?
Q9. Discuss the various drawbacks of floating point representation.
Q10. Discuss the various advantages of normalized floating point numbers.
Q11. Explain how real numbers and integers are stored in computer memory?
Q12. Discuss underflow and overflow conditions while performing normalized arithmetic
operations.
Q13. What are errors? Discuss the various types of errors that occur while performing numerical
computations. What measures can be taken to improve accuracy in numerical computations?
Q14. Define the term absolute error and relative error.
Q15. Differentiate between absolute error and relative error. Which is a better measure and
why? Give examples to support your answer.
Q16. Explain primary sources of errors in numerical computation. What are various measures
that can be taken to eliminate or reduce such errors.
Q17. Differentiate between truncation error, round off error, absolute error and relative error, by
taking examples.
Q18. An approximate value of PI is given by 3.1428571 and its true value is 3.1415926. Find the
absolute and relative errors.
Q19. What is the error in density of a cube when its mass is uncertain by +-2% and length of its
edge is uncertain by +-1%?
Q20. If x=3.26426, find absolute, relative and percentage error if
a. x is truncated to 4 decimal places
b. x is rounded off to 4 decimal places
Q21. Explain mantissa.
Q22. Explain with the help of suitable example that with normalized floating point
representation, following laws are not always valid:
a. Associative law: (a+b)+c=a+(b+c)
b. Distributive law: a(b+c)=ab+ac
Q23. Comment on the results:
a. Add 0.5467E99 to 0.7254E99
b. Add 0.4273E-2 and 0.5324E-3
c. Subtract 0.7253E2 from 0.5467E5
d. Subtract 0.3215E5 from 0.2312E6
Q24. Comment on the results:
a. Multiply 0.6543E5 by 0.2255E3
b. Multiply 0.2345E5 by 0.4201E3
c. Multiply 0.1246E-5 by 0.4116E-5
d. Divide 0.9998E5 by 0.1000E-99
Q25. Using normalized floating point arithmetic, determine the value of (x^2-y^2)/(x+y) when
x=0.4845 and y=0.4800. Suggest the precautions to be taken to prevent the errors and pitfalls
inherent in the floating point method.
Q26. For x=0.4845 and y=0.4800. Calculate the value of (x^2-y^2)/(x+y) by using normalized
floating point arithmetic. Compare the result with the value of (x-y). Indicate the error in the
former.
Q27. Assuming that the mantissas are truncated to 4 decimal digits, compute the error in the
following computations if any:
a. 6.7789 - 2.4523
b. 7.66689 - 1.45327
c. 6.7789 + 2.4523
d. 7.66689 + 1.45327
