You are on page 1of 7

ELE335 Paper (2010)

Page 1

Question 1 A 9-bit floating point number in binary representation may take the form seeeemmmm where s is the sign bit, eeee signifies the exponent bits and mmmm signifies the mantissa bits. a) If the exponent bits are in the form of an excess-7 code, explain the meaning of the three parameters. Illustrate your answer by calculating the floating point representation for 11.510. [8 marks] A = 010010110, B = 110101001, C = 011011100 and D = 000000000 b) The numbers A, B, C and D are in the above floating point format (including the excess-7 code). i) In the same binary floating point form (i.e. do not convert to decimal then back to perform calculations), calculate the value of: (A B) + C ii) Comment on the accuracy of your answer [12 marks] c) If this floating point format handles very small and very large values by using the same special cases as the IEEE-754 standard, what floating point value should you get if you divide the result of part (b) by D and why? [5 marks]

Question 2 A 9-bit floating point number in binary representation may take the form seeeemmmm where s is the sign bit, eeee signifies the exponent bits and mmmm signifies the mantissa bits. a) If the exponent bits are in the form of an excess-7 code, explain the meaning of the three parameters. Illustrate your answer by calculating the floating point representation for 0.3437510. [8 marks] A = 010001010, B = 010111100, C = 110001000 b) The numbers A, B, C and D are in the above floating point format (including the excess-7 code). iii)In the same binary floating point form (i.e. do not convert to decimal then back to perform calculations), calculate the value of: (A B) - C iv) Comment on the accuracy of your answer [12 marks] c) If this floating point format handles very small and very large values by using the same special cases as the IEEE-754 standard, show how the value 1.00112 2-8 can be represented. Comment on the issues this raises in terms of accuracy. [5 marks]

Question 3 A 9-bit floating point number in binary representation may take the form seeeemmmm where s is the sign bit, eeee signifies the exponent bits and mmmm signifies the mantissa bits. a) If the exponent bits are in the form of an excess-7 code, explain the meaning of the three parameters. Illustrate your answer by calculating the floating point representation for 0.2187510. [8 marks] A = 010001010, B = 010111100 b) The numbers A and B are in the above floating point format (including the excess-7 code). i) In the same binary floating point form (i.e. do not convert to decimal then back to perform calculations), calculate the value of:
A2 +B 2

ii) Comment on the accuracy of your answer [12 marks] c) If this floating point format handles very small and very large values by using the same special cases as the IEEE-754 standard, what are the largest value and the smallest non-zero value which may be represented? [5 marks]

Question 4 a) A 9-bit floating point number in binary representation may take the form seeeemmmm where s is the sign bit, eeee signifies excess-7 coded exponent bits and mmmm signifies the mantissa bits. Explain the meaning of these three parameters and describe the process you would use to multiply two such floating point values together. [6marks] b) A 9-bit floating point number representation is described in part (a). Using only the components

Figure 2: Digital components: a) Array Multiplier b) Shifter c) Adder d) Subtractor shown in Figure 2 plus one XOR gate, design a circuit that can perform the multiplication of two such floating point numbers in hardware. Your circuit should give a result in the same floating point format. [12 marks] A = 010001010, B = 010111100 c) The numbers A and B are in the above floating point format (including the excess-7 code). i) In the same binary floating point form (i.e. do not convert to decimal then back to perform calculations), calculate the value of:

+B

[7 marks]

ii) Comment on the accuracy of your answer

Question 5

Figure 2: Digital components: a) Array Multiplier b) Shifter c) Adder d) Subtractor a) Showing all your working, perform the following binary multiplication: 1.10112 1.01012 [5 marks]

b) Figure 2(a) shows a 5-bit array multiplier for multiplying two positive binary values A and B together (i.e. it does not correct for 2's complement values). It has two inputs A4...A0 and B4...B0 and one output M9...M0. Between which two bits of the output should the binary point be if the device is used to perform the multiplication from part (a)? Explain your answer. [2 marks]

c) A 9-bit floating point number in binary representation may take the form seeeemmmm where s is the sign bit, eeee signifies excess-7 coded exponent bits and mmmm signifies the mantissa bits. Explain the meaning of these three parameters and describe the process you would use to multiply two such floating point values together. [6marks]

d) A 9-bit floating point number representation is described in part (c). Using only the components shown in Figure 2 plus one XOR gate, design a circuit that can perform the multiplication of two such floating point numbers in hardware. Your circuit should give a result in the same floating point format. [12 marks]

Question 6 a) A 7-bit floating point number is of the form seeemmm where s stands for the sign bit, e stands for an exponent bit and m stands for a mantissa bit. ii) Explain the significance of the exponent bits. Explain why, in this case, would they normally be in excess-3 form. iii)If the mantissa bits are 010, explain the value of the mantissa. [6 marks]

i) Explain the meaning of the sign bit.

b) Two numbers in the above floating point form are: A = 0011011 and B = 0010100 Calculate the product (i.e. A x B) of these two numbers in the same floating point form; do not convert to decimal then back to binary floating point. Show your working. [6 marks] c) Two numbers in the above floating point form are: C = 0111100 and D = 1110110 Calculate the sum (i.e. C + D) of these two numbers in the same floating point form; do not convert to decimal then back to binary floating point. Show your working. [7 marks] d) Convert octal number 138 into binary form. Negate the result to give an 8 bit, 2s complement form. Show your working. [6 marks]

You might also like