This problem concerns the design of a divider for floating-point numbers: (F 1 x 2 E1 )/(F
Question:
This problem concerns the design of a divider for floating-point numbers:
(F1 x 2E1)/(F2 x 2E2) = F x 2E
Assume that F1 and F2 are properly normalized fractions (or 0), with negative fractions expressed in 2’s complement. The exponents are integers with negative numbers expressed in 2’s complement. The result should be properly normalized if it is not zero. Fractions are 8 bits including sign, and exponents are 5 bits including sign.
(a) Draw a flowchart for the floating-point divider. Assume that a divider is available that will divide two binary fractions to give a fraction as a result. Do not show the individual steps in the division of the fractions on your flowchart; just say “divide.” The divider requires that |F2| > |F1| before division is carried out.
(b) Illustrate your procedure by computing 0.111 x 23/1.011 x 2-2 When you divide F1 by F2, you don’t need to show the individual steps, just the result of the division.
(c) Write a Verilog description for the system.
Step by Step Answer:
Digital Systems Design Using Verilog
ISBN: 978-1285051079
1st edition
Authors: Charles Roth, Lizy K. John, Byeong Kil Lee