WorksheetsL6 Data Rep
Total questions: 15
Worksheet time: 8mins
Which is the largest normalized FP value (5-bit M, 4-bit E)?
00111 0111
01111 0110
11111 0111
10111 1111
Which is the smallest normalized FP value (5-bit M, 4-bit E)?
01000 1000
00001 1000
11111 1111
01000 1111
How do you negate a 2s complement number?
Invert bits and add one
Invert bits
Add one and invert bits
Write it backwards
How do you perform a - b in signed binary?
a + (-b)
a + (-1)*b
a - |b|
a + (b^-1)
What is the absolute error of the representation of 0.7501 in fixed point binary?
-0.0001
0.0001
0.2501
0.2499
What is the relative error of a binary representation of a number?
absolute error / represented number (as a %)
absolute error / intended number (as a %)
represented number / absolute error (as a %)
intended number / absolute error (as a %)
Which signed binary number is most negative?
10111
11011
11101
11110
Which signed binary number is the closest to zero?
10111
11011
11101
11110
Divide this signed binary byte by 2: 10010010
(a)
What s__ d__ is another name for the mantissa in a floating point number?
(a)
If the mantissa is M and the exponent E, what is the value of a binary floating point number?
E * 2^M
M * 2^E
E * 10^M
M * 10^E
Which of these floating point numbers is not normalised?
1100 1000
1000 0011
0110 0111
0000 1111
Convert 0.345 to 6-bit fixed point binary using the doubling algorithm
0.01011
0.01010
0.01101
0.01110
Why is relative error more useful than absolute error in a binary representation?
Because it gives an indication of the impact/effect of the error.
Because it gives an indication of the scale of the error.
Because it gives an indication of the size of the error.
Because it gives an indication of the cause of the error.
What is the normalized floating point representation of 1664 with 5 bits for the mantissa and 5 bits for the exponent. Enter M and E as 2 binary values with a space between them.
(a)
